Step 8.2_Aanlysis by models.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#China\n",
    "Country='China'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "data=pd.read_csv('/Volumes/BECCA/Project/Data/Result/%s/By_model/Total_Result.csv' %Country, index_col=[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>5942</td>\n",
       "      <td>971</td>\n",
       "      <td>1327</td>\n",
       "      <td>619</td>\n",
       "      <td>0.820522</td>\n",
       "      <td>0.905655</td>\n",
       "      <td>0.859540</td>\n",
       "      <td>0.881995</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>5932</td>\n",
       "      <td>1031</td>\n",
       "      <td>1267</td>\n",
       "      <td>629</td>\n",
       "      <td>0.812620</td>\n",
       "      <td>0.904130</td>\n",
       "      <td>0.851932</td>\n",
       "      <td>0.877255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>kNN</td>\n",
       "      <td>5894</td>\n",
       "      <td>1220</td>\n",
       "      <td>1078</td>\n",
       "      <td>667</td>\n",
       "      <td>0.786996</td>\n",
       "      <td>0.898339</td>\n",
       "      <td>0.828507</td>\n",
       "      <td>0.862011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Tree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>5904</td>\n",
       "      <td>1188</td>\n",
       "      <td>1110</td>\n",
       "      <td>657</td>\n",
       "      <td>0.791737</td>\n",
       "      <td>0.899863</td>\n",
       "      <td>0.832487</td>\n",
       "      <td>0.864865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Tree</td>\n",
       "      <td>gnb</td>\n",
       "      <td>5979</td>\n",
       "      <td>1514</td>\n",
       "      <td>784</td>\n",
       "      <td>582</td>\n",
       "      <td>0.763404</td>\n",
       "      <td>0.911294</td>\n",
       "      <td>0.797945</td>\n",
       "      <td>0.850861</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>k3</td>\n",
       "      <td>dtree</td>\n",
       "      <td>6096</td>\n",
       "      <td>1559</td>\n",
       "      <td>739</td>\n",
       "      <td>465</td>\n",
       "      <td>0.771532</td>\n",
       "      <td>0.929127</td>\n",
       "      <td>0.796342</td>\n",
       "      <td>0.857625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>k2</td>\n",
       "      <td>dtree</td>\n",
       "      <td>6311</td>\n",
       "      <td>1789</td>\n",
       "      <td>509</td>\n",
       "      <td>250</td>\n",
       "      <td>0.769839</td>\n",
       "      <td>0.961896</td>\n",
       "      <td>0.779136</td>\n",
       "      <td>0.860924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Tree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>6257</td>\n",
       "      <td>1686</td>\n",
       "      <td>612</td>\n",
       "      <td>304</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.953666</td>\n",
       "      <td>0.787738</td>\n",
       "      <td>0.862796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>k3</td>\n",
       "      <td>SVM</td>\n",
       "      <td>6096</td>\n",
       "      <td>1559</td>\n",
       "      <td>739</td>\n",
       "      <td>465</td>\n",
       "      <td>0.771532</td>\n",
       "      <td>0.929127</td>\n",
       "      <td>0.796342</td>\n",
       "      <td>0.857625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>k2</td>\n",
       "      <td>SVM</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>6069</td>\n",
       "      <td>1506</td>\n",
       "      <td>792</td>\n",
       "      <td>492</td>\n",
       "      <td>0.774467</td>\n",
       "      <td>0.925011</td>\n",
       "      <td>0.801188</td>\n",
       "      <td>0.858659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>6050</td>\n",
       "      <td>1485</td>\n",
       "      <td>813</td>\n",
       "      <td>511</td>\n",
       "      <td>0.774692</td>\n",
       "      <td>0.922116</td>\n",
       "      <td>0.802920</td>\n",
       "      <td>0.858400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>5980</td>\n",
       "      <td>1427</td>\n",
       "      <td>871</td>\n",
       "      <td>581</td>\n",
       "      <td>0.773338</td>\n",
       "      <td>0.911446</td>\n",
       "      <td>0.807344</td>\n",
       "      <td>0.856243</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>k3</td>\n",
       "      <td>nn</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>k2</td>\n",
       "      <td>nn</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>k5</td>\n",
       "      <td>lr</td>\n",
       "      <td>6019</td>\n",
       "      <td>1448</td>\n",
       "      <td>850</td>\n",
       "      <td>542</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.917391</td>\n",
       "      <td>0.806080</td>\n",
       "      <td>0.858141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>k3</td>\n",
       "      <td>lr</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>k2</td>\n",
       "      <td>lr</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>k5</td>\n",
       "      <td>sgd</td>\n",
       "      <td>6019</td>\n",
       "      <td>1448</td>\n",
       "      <td>850</td>\n",
       "      <td>542</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.917391</td>\n",
       "      <td>0.806080</td>\n",
       "      <td>0.858141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>k3</td>\n",
       "      <td>sgd</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>k2</td>\n",
       "      <td>sgd</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Model Classifier    TP    FP    TN   FN  Accuracy    Recall  Precision  \\\n",
       "0     k7        kNN  5942   971  1327  619  0.820522  0.905655   0.859540   \n",
       "1     k6        kNN  5932  1031  1267  629  0.812620  0.904130   0.851932   \n",
       "2     k5        kNN  5894  1220  1078  667  0.786996  0.898339   0.828507   \n",
       "3   Tree        kNN  5904  1188  1110  657  0.791737  0.899863   0.832487   \n",
       "4   Tree        gnb  5979  1514   784  582  0.763404  0.911294   0.797945   \n",
       "5     k3      dtree  6096  1559   739  465  0.771532  0.929127   0.796342   \n",
       "6     k2      dtree  6311  1789   509  250  0.769839  0.961896   0.779136   \n",
       "7   Tree      dtree  6257  1686   612  304  0.775370  0.953666   0.787738   \n",
       "8     k3        SVM  6096  1559   739  465  0.771532  0.929127   0.796342   \n",
       "9     k2        SVM  6081  1557   741  480  0.770064  0.926840   0.796151   \n",
       "10    k6         nn  6069  1506   792  492  0.774467  0.925011   0.801188   \n",
       "11    k5         nn  6050  1485   813  511  0.774692  0.922116   0.802920   \n",
       "12    k4         nn  5980  1427   871  581  0.773338  0.911446   0.807344   \n",
       "13    k3         nn  6006  1459   839  555  0.772661  0.915409   0.804555   \n",
       "14    k2         nn  6081  1557   741  480  0.770064  0.926840   0.796151   \n",
       "15    k5         lr  6019  1448   850  542  0.775370  0.917391   0.806080   \n",
       "16    k3         lr  6006  1459   839  555  0.772661  0.915409   0.804555   \n",
       "17    k2         lr  6081  1557   741  480  0.770064  0.926840   0.796151   \n",
       "18    k5        sgd  6019  1448   850  542  0.775370  0.917391   0.806080   \n",
       "19    k3        sgd  6006  1459   839  555  0.772661  0.915409   0.804555   \n",
       "20    k2        sgd  6081  1557   741  480  0.770064  0.926840   0.796151   \n",
       "\n",
       "    F1_Score  \n",
       "0   0.881995  \n",
       "1   0.877255  \n",
       "2   0.862011  \n",
       "3   0.864865  \n",
       "4   0.850861  \n",
       "5   0.857625  \n",
       "6   0.860924  \n",
       "7   0.862796  \n",
       "8   0.857625  \n",
       "9   0.856539  \n",
       "10  0.858659  \n",
       "11  0.858400  \n",
       "12  0.856243  \n",
       "13  0.856410  \n",
       "14  0.856539  \n",
       "15  0.858141  \n",
       "16  0.856410  \n",
       "17  0.856539  \n",
       "18  0.858141  \n",
       "19  0.856410  \n",
       "20  0.856539  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# By feature"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #All"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k7'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kNN</td>\n",
       "      <td>5942</td>\n",
       "      <td>971</td>\n",
       "      <td>1327</td>\n",
       "      <td>619</td>\n",
       "      <td>0.820522</td>\n",
       "      <td>0.905655</td>\n",
       "      <td>0.85954</td>\n",
       "      <td>0.881995</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Classifier    TP   FP    TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "0        kNN  5942  971  1327  619  0.820522  0.905655    0.85954  0.881995"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TP   FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7  kNN  kNN  kNN  kNN      kNN    kNN       kNN      kNN"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=pd.DataFrame(opt, index=[0])\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=6"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k6'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>kNN</td>\n",
       "      <td>5932</td>\n",
       "      <td>1031</td>\n",
       "      <td>1267</td>\n",
       "      <td>629</td>\n",
       "      <td>0.812620</td>\n",
       "      <td>0.904130</td>\n",
       "      <td>0.851932</td>\n",
       "      <td>0.877255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>nn</td>\n",
       "      <td>6069</td>\n",
       "      <td>1506</td>\n",
       "      <td>792</td>\n",
       "      <td>492</td>\n",
       "      <td>0.774467</td>\n",
       "      <td>0.925011</td>\n",
       "      <td>0.801188</td>\n",
       "      <td>0.858659</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "1         kNN  5932  1031  1267  629  0.812620  0.904130   0.851932  0.877255\n",
       "10         nn  6069  1506   792  492  0.774467  0.925011   0.801188  0.858659"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TP   FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7  kNN  kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6   nn   nn  kNN  kNN      kNN     nn       kNN      kNN"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k5'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>kNN</td>\n",
       "      <td>5894</td>\n",
       "      <td>1220</td>\n",
       "      <td>1078</td>\n",
       "      <td>667</td>\n",
       "      <td>0.786996</td>\n",
       "      <td>0.898339</td>\n",
       "      <td>0.828507</td>\n",
       "      <td>0.862011</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>nn</td>\n",
       "      <td>6050</td>\n",
       "      <td>1485</td>\n",
       "      <td>813</td>\n",
       "      <td>511</td>\n",
       "      <td>0.774692</td>\n",
       "      <td>0.922116</td>\n",
       "      <td>0.802920</td>\n",
       "      <td>0.858400</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>lr</td>\n",
       "      <td>6019</td>\n",
       "      <td>1448</td>\n",
       "      <td>850</td>\n",
       "      <td>542</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.917391</td>\n",
       "      <td>0.806080</td>\n",
       "      <td>0.858141</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>sgd</td>\n",
       "      <td>6019</td>\n",
       "      <td>1448</td>\n",
       "      <td>850</td>\n",
       "      <td>542</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.917391</td>\n",
       "      <td>0.806080</td>\n",
       "      <td>0.858141</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "2         kNN  5894  1220  1078  667  0.786996  0.898339   0.828507  0.862011\n",
       "11         nn  6050  1485   813  511  0.774692  0.922116   0.802920  0.858400\n",
       "15         lr  6019  1448   850  542  0.775370  0.917391   0.806080  0.858141\n",
       "18        sgd  6019  1448   850  542  0.775370  0.917391   0.806080  0.858141"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeoAAAEvCAYAAACKSII9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nO3de3hV1Z3/8fc3Jwm5QZCAKIRAVFAQDGhEUAQhotRaLTqOt94cpzq1YqvtdHC01lodW9tOR5/SztjWwar1Un/VcTq2UhLRar0QvCIgAkoIF7kpBsj9fH9/7JPkJCThICFnJ/m8nocnZ++19j5rH2M+Z+299l7m7oiIiEg4pSS7ASIiItIxBbWIiEiIKahFRERCTEEtIiISYgpqERGREFNQi4iIhFhqshvQ1uDBg33UqFHJboaIiEi3WbZs2XZ3H9JeWeiCetSoUZSXlye7GSIiIt3GzNZ3VKZT3yIiIiGmoBYREQkxBbWIiEiIhe4atUhvVV9fT2VlJTU1NcluSo+VkZFBfn4+aWlpyW6KSLdRUIt0k8rKSvr378+oUaMws2Q3p8dxd3bs2EFlZSWFhYXJbo5It9Gpb5FuUlNTQ15enkL6UzIz8vLydEZC+hwFtUg3UkgfHH1+0hcpqEX6mCeeeAIzY9WqVcluiogkQEEt0sc8/PDDTJs2jUceeeSQvUdjY+Mh27dIX6Oglr6pvgbWlMIbv4OVf4T3n4dNb8DOdbBnBzTWJ7uFh8Tu3bt58cUX+c1vftMqqO+66y4mTJhAUVER8+fPB2DNmjWceeaZFBUVceKJJ7J27VqWLFnCueee27zdtddey8KFC4HgqYK33XYb06ZN4/e//z2/+tWvOPnkkykqKuLCCy9k7969AHz44YfMnTuXoqIiioqK+Nvf/sZ3v/td7r777ub93nTTTdxzzz3d8ImIhF9Co77NbA5wNxABfu3uP2xTPhK4DxgC7AS+4O6VZjYR+CUwAGgE7nD3R7uw/SKJ+2QTvLcIVi+Cdc9C/d7O66dmQsYA6Dcg9rN/3OvcNmXxP3NbllP7dc+xJejJJ59kzpw5jBkzhkGDBvHaa6/x4Ycf8uSTT/LKK6+QlZXFzp07Abj88suZP38+c+fOpaamhmg0yoYNGzrdf0ZGBi+88AIAO3bs4Ktf/SoAN998M7/5zW+YN28e1113HTNmzOCJJ56gsbGR3bt3M2zYMC644AK+8Y1vEI1GeeSRR3j11VcP7Ych0kPsN6jNLAIsAGYDlcBSM3vK3VfEVfsJ8Ft3v9/MZgF3Al8E9gJfcvf3zGwYsMzMnnH3j7v8SETaikZh02uw+s+w+hnY8lawPrcAJl4GY+ZA3tFQWwU1n0DtJ21+7tp3/SebW5br9+y/DZF+LQE+5aewPRVSIny/dCsrttUAFvwz9n1tseUEjRs2gO997vhO6zz88MN885vfBOCSSy7h4YcfJhqNcsUVV5CVlQXAoEGDqKqqYuPGjcydOxcIAjgRF198cfPr5cuXc/PNN/Pxxx+ze/duzj77bADKysr47W9/C0AkEiE3N5fc3Fzy8vJ4/fXX+fDDD5k0aRJ5eXkJH7tIb5ZIj3oysMbd1wGY2SPA+UB8UI8Dro+9fhZ4EsDdVzdVcPdNZraVoNetoJZDo2YXrC0Les3vLYK928FSYMQUOPPWIJyHHBcLwYPU2BCE9j4B30HQWwTcoaEWGusO8PS6xbU57nVzmFtwOr/6o+B9UiKtf5qxY+dOysrKWL58OWZGY2MjZsaFF164z2hqd2+3FampqUSj0ebltrdKZWdnN7/+yle+wpNPPklRURELFy5kyZIlnR7hP/7jP7Jw4UK2bNnCP/zDPyTyoYj0CYkE9XAg/nxXJXBKmzpvAhcSnB6fC/Q3szx339FUwcwmA+nA2oNqsUg8d9ixpqXXXPESRBsg8zA4ZjaMORuOngVZg7r+vSOpwX4T3ffKlTBkDADfu2RsS/u9EaKNsZ/RNstxP9sti61r8tEHHby58fiDf+BLF32O//rpbc0hPuNzlzIoK8J9v/pPLjuvhKzs/uz8+BMG5Q0mf/gwnnz8MT4/dy61dfU0Rp2RI0eyYsUKamtrqampobS0lGnTprX7jlVVVRx55JHU19fz0EMPMXz4cABKSkr45S9/yTe/+U0aGxvZs2cPAwYMYO7cudxyyy3U19fzu9/9LrHPVKQPSCSo2+t6tP26/W3g52b2FeB5YCPQ0LwDsyOBB4Avu3u0zbaY2VXAVQAFBQUJNVz6sIZaWP9iEMyrn4GP3g/WH348nDov6DUPLw6CNOzMwFIh5SDa2hz2nYS8R3n4qb8w/xv/BClpwfqGWi48ZxYrV6/lvJJTKZ5WQnpaGufMOo1/u3EeD/z0Jq7+l9u55ZabSEtN5ff/9WOOGlXA3392FiccfxyjjxrFpOPHwJ7tsKsy2OfurZAJpET4wS3/yimTJzNyZAETJoynqmoPuHP33Xdz1VVX8Zvf/IZIJMIvf/lLpk6dSnp6OjNnzmTgwIFEIpEu+4hFejrr6BRXcwWzqcCt7n52bPlGAHe/s4P6OcAqd8+PLQ8AlgB3uvvv99eg4uJi13zUso+qLbGBYM/AuiVQtxtSM6BwBow5C0afDQNHJLuVnVq5ciVjx45NdjPa5x70ztsN+c6+BLRZn4h9Ts2nQEqEqBsnnnEOv7//14weMzq2PrW5vKn+yndXh/dzFPmUzGyZuxe3V5bI1/ilwGgzKyToKV8CXNbmDQYDO2O95RsJRoBjZunAEwQDzfYb0iLNolHY/HpLr3nzG8H6Aflwwt8HveZRp0N6VnLb2VuYtQTnp9Vu2EeDSxEdfgmIQmM9K955h3O/eA1z58xk9NAM2NXJ6PKPt8EPz97PyPt2RuDH/0zPgRTdnSo9w36D2t0bzOxa4BmC27Puc/d3zOw2oNzdnwLOAO40Myc49f312OZ/D0wH8mKnxQG+4u5vdO1hSK9Q80lw21TTQLA9W4PeVP5kKLklCOfDx3XNQDDpegcR9uMOP451689rCftWQd8m4DNq4ISLWw/U+2Qj1KxsWd5v7972E/ADID2bAxl1L33MuM/D4GO65a0SujDm7k8DT7dZd0vc68eBx9vZ7kHgwYNso/RmO9a2DARb/zeI1ge9oGPODIL5mDMPzUAwCaemsCcSdAvak7ETzvlxx/twD+6R39+tdm1/Vm2G7e+2LEcbOn4PkcPHhiuoRbpMQx1U/C3oNa/+M+yM3QQwZCxMvSYI5/zJPWMgmISTWdAbTs8Gjvx0+3DvtU+nky5yMANAD5D+Gsqht3try0Cwtc9CXVXwIJDC6TDlazD6LDhsZLJbKdLCDFLTk90KEUBBLYdCNApb3mwZCLbptWB9/2Ew4cKg11w4PdbjERGRziiopWvUVgW3Ta1+Jug97/4QMMg/GWbdHNw+dcQEDQRLskgkwoQJE2hoaKCwsJAHHniAgQMHdtn+Fy5cSHl5OT//+c+59dZbycnJ4dvf/naX7V+kL1JQy6e3c11Lr3n9i8FjMfvlwjGzWgaCZQ9OdislTmZmJm+8Edx08eUvf5kFCxZw0003JblVItIZBbUkrrE+eERnUzjveC9YP/hYOOXqIJxHnAKRtOS2UxIydepU3nrrreblH//4xzz22GPU1tYyd+5cvv/97wPw29/+lp/85CeYGSeccAIPPPAA//u//8vtt99OXV0deXl5PPTQQwwdOjRZhyLSqymopXO7t8Gav8QGgpUFt61E0mHUNJj81WAg2KDCZLdSDlBjYyOlpaVceeWVACxatIj33nuPV199FXfnvPPO4/nnnycvL4877riDF198kcGDBzdPgTlt2jRefvllzIxf//rX3HXXXfz0pz9N5iGJ9FoKamnNPZgOsqnXvHEZ4JBzBBz/+dhAsBnQLyfZLe3Z/jQftrzdtfs8YgJ85oedVqmurmbixIl88MEHnHTSScyePRsIgnrRokVMmjQJgN27d/Pee+/x5ptv8nd/93cMHhxcwhg0KLinvbKykosvvpjNmzdTV1dHYaG+rIkcKgpqgbo9rQeCVW0GDIafBDP/Neg1H1mkgWC9QNM16l27dnHuueeyYMECrrvuOtydG2+8kauvvrpV/XvuuWefKTAB5s2bxw033MB5553HkiVLuPXWW7vpCET6HgV1X7Xz/ZZ7mz94ARprg0cnHj0zNhBsNuQMSXYre6/99HwPtdzcXO655x7OP/98vva1r3H22Wfz3e9+l8svv5ycnBw2btxIWloaJSUlzJ07l+uvv568vDx27tzJoEGD2LVrV/O0lffff39Sj0Wkt1NQ9xWN9bDhlZZT2tvfDdbnjQ6uNY85G0ZM0UMe+pBJkyZRVFTEI488whe/+EVWrlzJ1KlTAcjJyeHBBx/k+OOP56abbmLGjBlEIhEmTZrEwoULufXWW7nooosYPnw4U6ZM4f3330/y0Yj0Xvud5rK7aZrLLrRnR8tAsDWlwfOOU9Jg1GlBr3n0WZB3dLJb2WeEeprLHkSfo/RGBzvNpfQU7vDh8pZec+VSgoFgQ2Hc54JwPuoM6Nc/yQ0VEZFEKah7urq98P5zLQPBPtkYrB82Cc6YHxsINlFz74qI9FAK6p7oo/VxA8H+Cg01kJ4TDAQ748YgnPvr4RMiIr2BgronaGyAyldbTmlvWxmsH3QUFP9DEMwjT4XUfsltp4iIdDkFdVjt3QlrFscGgi2Gmo+D+U9HngqTvhBcb+6mSctFRCR5FNRh4Q5bV8QNBHsVPArZQ+C4zwa95qNnQkZuslsqIiLdSEGdTPXV8P7zLQPBdm0I1h9ZBNP/OZgactgkDQSTLhM/zeXYsWO5//77ycrKOqh9lpeX89vf/pZ77rmn3fJNmzZx3XXX8fjjjx/U+4j0VbqPurvtqmzpNb//PDRUQ1p20FsefVbwb8CRyW6lHAJhuP83JyeH3bt3A3D55Zdz0kknccMNNzSXuzvuTkqIvxyG4XMU6Wqd3Ucd3v8be4toI1S8DIu/D784FX52PPzfDcGTwU76MnzhD/Av78MlDwXLCmnpJqeffjpr1qzhgw8+YOzYsVxzzTWceOKJbNiwgUWLFjF16lROPPFELrroouZwX7p0KaeeeipFRUVMnjyZqqoqlixZwrnnngvAc889x8SJE5k4cSKTJk2iqqqKDz74gPHjxwNQU1PDFVdcwYQJE5g0aRLPPvssAAsXLuSCCy5gzpw5jB49mu985zvJ+VBEQkinvg+F6o+CJ4E1DQSr3hkMBCuYCrN/EBsINlqTXEjSNDQ08Kc//Yk5c+YA8O677/Lf//3f/OIXv2D79u3cfvvtLF68mOzsbH70ox/x7//+78yfP5+LL76YRx99lJNPPplPPvmEzMzMVvv9yU9+woIFCzjttNPYvXs3GRkZrcoXLFgAwNtvv82qVas466yzWL16NQBvvPEGr7/+Ov369ePYY49l3rx5jBgxohs+DZFwU1B3BXfYtqrllPaGV8AbISsveIb26LPg6FmQOTDZLZWQ+NGrP2LVzlVdus/jBh3Hv0z+l07rNE1zCUGP+sorr2TTpk2MHDmSKVOmAPDyyy+zYsUKTjvtNADq6uqYOnUq7777LkceeSQnn3wyAAMGDNhn/6eddho33HADl19+ORdccAH5+fmtyl944QXmzZsXtPe44xg5cmRzUJeUlJCbGwyWHDduHOvXr1dQi6Cg/vTqa4KHjTSF866KYP0RE+D0G4KBYMNPhJRIctspEqdpmsu2srOzm1+7O7Nnz+bhhx9uVeett95qd8rLePPnz+ezn/0sTz/9NFOmTGHx4sWtetWdjYnp16/lOQCRSISGhob9Ho9IX6CgPhC7NrY8Eez956B+L6RlBc/PPv2GoOecOzzZrZQeYH8932SaMmUKX//611mzZg3HHHMMe/fupbKykuOOO45NmzaxdOlSTj75ZKqqqvY59b127VomTJjAhAkTeOmll1i1alVzDx5g+vTpPPTQQ8yaNYvVq1dTUVHBsccey2uvvdbdhynSYyQU1GY2B7gbiAC/dvcftikfCdwHDAF2Al9w98pY2ZeBm2NVb3f3njN5bbQRNi5r6TV/+HawfmBB8NCR0WfDqGmQltH5fkR6kCFDhrBw4UIuvfRSamtrAbj99tsZM2YMjz76KPPmzaO6uprMzEwWL17catv/+I//4NlnnyUSiTBu3Dg+85nPsHnz5ubya665hn/6p39iwoQJpKamsnDhwlY9aRHZ135vzzKzCLAamA1UAkuBS919RVyd3wN/dPf7zWwWcIW7f9HMBgHlQDHgwDLgJHf/qKP3S/rtWdUfw9pSWL0omCJy7w6wCBRMCXrMY+bAkGM1EEwOmG4r6hr6HKU3OthpLicDa9x9XWxnjwDnAyvi6owDro+9fhZ4Mvb6bOAv7r4ztu1fgDlA64tfyeQO21e39JorXgoGgmUe1nJf8zElwbKIiEg3SySohwMb4pYrgVPa1HkTuJDg9PhcoL+Z5XWw7T4Xcc3sKuAqgIKCgkTb/unV18D6F4Je83vPwEcfBOuHjofTvhH0mvOLNRBMRESSLpGgbu8cb9vz5d8Gfm5mXwGeBzYCDQlui7vfC9wLwanvBNp04D7Z3DIQbN0SqN8DqZlw1Aw49bqg5zxQt4KIiEi4JBLUlUB8guUDm+IruPsm4AIAM8sBLnT3XWZWCZzRZtslB9HeA1NbBS/eE/SaN78ZrMsdARMvDQaCFZ4OaZmd70NERCSJEgnqpcBoMysk6ClfAlwWX8HMBgM73T0K3EgwAhzgGeDfzKzpAu9ZsfLukZoBS38FQ46Dku8Fp7QPH6uBYCIi0mPsN6jdvcHMriUI3Qhwn7u/Y2a3AeXu/hRBr/lOM3OCU99fj22708x+QBD2ALc1DSzrFpE0uGGles0iItJjJXQftbs/DTzdZt0tca8fB9qdw87d76Olh939FNIiItKDafYskT4kEok0z241ceJEPvjgA3bs2MHMmTPJycnh2muv3e8+7rvvPiZMmMAJJ5zA+PHj+Z//+Z9uaLlI36VHiIr0Ie0963vPnj384Ac/YPny5SxfvrzT7SsrK7njjjt47bXXyM3NZffu3Wzbtu2g2tTY2EgkolshRTqiHrVIH5ednc20adP2mZKyPVu3bqV///7k5OQAkJOTQ2FhIQBr1qzhzDPPpKioiBNPPJG1a9fi7vzzP/8z48ePZ8KECTz66KMALFmyhJkzZ3LZZZcxYcIEAB588EEmT57MxIkTufrqq2lsbDxERyzSs6hHLZIEW/7t36hd2bXTXPYbexxH/Ou/dlonfprLwsJCnnjiiQN6j6KiIoYOHUphYSElJSVccMEFfO5znwPg8ssvZ/78+cydO5eamhqi0Sh/+MMfeOONN3jzzTfZvn07J598MtOnTwfg1VdfZfny5RQWFrJy5UoeffRRXnzxRdLS0rjmmmt46KGH+NKXvvQpPgmR3kVBLdKHdDTNZaIikQh//vOfWbp0KaWlpVx//fUsW7aMb33rW2zcuJG5c+cCNPfOX3jhBS699FIikQhDhw5lxowZLF26lAEDBjB58uTm3nhpaSnLli1rnuu6urqaww8//CCPVqR3UFCLJMH+er5hZmZMnjyZyZMnM3v2bK644gpuuOGGdut2NulP2zmwv/zlL3PnnXd2eXtFejpdoxaRhG3atKnV3NFvvPEGI0eOZMCAAeTn5/Pkk8F8PLW1tezdu5fp06fz6KOP0tjYyLZt23j++eeZPHnyPvstKSnh8ccfZ+vWrQDs3LmT9evXd89BiYScetQiwqhRo/jkk0+oq6vjySefZNGiRYwbN26fevX19Xz7299m06ZNZGRkMGTIEP7zP/8TgAceeICrr76aW265hbS0NH7/+98zd+5cXnrpJYqKijAz7rrrLo444ghWrWp9fX7cuHHcfvvtnHXWWUSjUdLS0liwYAEjR47sluMXCbP9zkfd3ZI+H7XIIaJ5lLuGPkfpjTqbj1qnvkVEREJMp75FpF2nnHIKtbW1rdY98MADzfc9i0j3UFCLSLteeeWVZDdBRNCpb5FuFbYxIT2NPj/pixTUIt0kIyODHTt2KGw+JXdnx44dCT3qVKQ30alvkW6Sn59PZWXlQU9i0ZdlZGSQn5+f7GaIdCsFtUg3SUtLa35kpohIonTqW0REJMQU1CIiIiGmoBYREQkxBbWIiEiIKahFRERCTEEtIiISYgpqERGREFNQi4iIhFhCQW1mc8zsXTNbY2bz2ykvMLNnzex1M3vLzM6JrU8zs/vN7G0zW2lmN3b1AYiIiPRm+w1qM4sAC4DPAOOAS81sXJtqNwOPufsk4BLgF7H1FwH93H0CcBJwtZmN6pqmi4iI9H6J9KgnA2vcfZ271wGPAOe3qePAgNjrXGBT3PpsM0sFMoE64JODbrWIiEgfkUhQDwc2xC1XxtbFuxX4gplVAk8D82LrHwf2AJuBCuAn7r7zYBosIiLSlyQS1NbOurbz9F0KLHT3fOAc4AEzSyHojTcCw4BC4FtmdtQ+b2B2lZmVm1m5ZhYSERFpkUhQVwIj4pbzaTm13eRK4DEAd38JyAAGA5cBf3b3enffCrwIFLd9A3e/192L3b14yJAhB34UIiIivVQiQb0UGG1mhWaWTjBY7Kk2dSqAEgAzG0sQ1Nti62dZIBuYAqzqqsaLiIj0dvsNandvAK4FngFWEozufsfMbjOz82LVvgV81czeBB4GvuLuTjBaPAdYThD4/+3ubx2C4xAREemVLMjT8CguLvby8vJkN0NERKTbmNkyd9/n0jDoyWQiIiKhpqAWEREJMQW1iIhIiCmoRUREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEFNQiIiIhpqAWEREJMQW1iIhIiCmoRUREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEFNQiIiIhpqAWEREJMQW1iIhIiCmoRUREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEUpPdgEPJ3dl4/Q2kZGSQkpVFSlYmlpVFSmYWKZmZpGTHfmZlYZmZpGRlk5IVLKdkZmKZmZhZsg9DRET6sISC2szmAHcDEeDX7v7DNuUFwP3AwFid+e7+dKzsBOC/gAFAFDjZ3Wu67Ag6U19P7Zr38L3VRPfuJVpdjdfWJr69WRDgmS3hvU/gt13fFPidfRHIzMRSe/V3JBER6SL7TQsziwALgNlAJbDUzJ5y9xVx1W4GHnP3X5rZOOBpYJSZpQIPAl909zfNLA+o7/Kj6Kjt6ekc/cc/tlrnDQ1Ea2qI7tmLV+9tDvDo3r1EmwN9L15dTXRPXFmsbnPo79jZalvfu/eA25aSmYlldxD4+3wRiA/8WFnbLwLZ2Vhams4CiIj0Iol06yYDa9x9HYCZPQKcD8QHtRP0mAFygU2x12cBb7n7mwDuvqMrGn0wLDWVSE4OkZycLt2vu+M1NS3h/Sm/CDRs3dpmm73Q2Jh4QyKRVmcAOvoi0BzuTWVtvwhktd7GMjKwFA1pEBHpbokE9XBgQ9xyJXBKmzq3AovMbB6QDZwZWz8GcDN7BhgCPOLud7V9AzO7CrgKoKCg4EDaHxoWd5q8K7k7Xl+PdxT4cV8M2u39x9Y17tpF/eZNn/4yAMT15uN+ZmdhHV4CyEroi4AuA4iIdCyRv5DtnUf1NsuXAgvd/admNhV4wMzGx/Y/DTgZ2AuUmtkydy9ttTP3e4F7AYqLi9vuu08zMyw9HdLTiQwc2KX77vgyQJvA398XgR07W10CiFZXgyf+n7H5MkCbXnzLF4G46/77+SIQvx9LT9dlABHp8RIJ6kpgRNxyPi2ntptcCcwBcPeXzCwDGBzb9jl33w5gZk8DJwKlSNIl7zJAYl8EGrZtI7o+tj5WRkND4g1pugzQdPq+nS8Clp4GGFjTP4KfEAv5+LJY6FubMmi1vcXXbbvvWN199h2r2lzWwfYW/37t7Ltl27Zt62zfbY65g7Y1l7Wzfav9ttq+nX03lbd3zB3tO/4zjt932/222r6dNrfZfp99t2mbvuhJR1KPPLLL/3Z2+F4J1FkKjDazQmAjcAlwWZs6FUAJsNDMxgIZwDbgGeA7ZpYF1AEzgJ91UdslpA7VZQAAr6vr+Lp/gl8EGj/5hPotm4Pgr2/A8eAckXvLmYCm1+7B6aN21jet8w7Wt9q2ve1FpMfKX/Bz+peUdMt77Teo3b3BzK4lCN0IcJ+7v2NmtwHl7v4U8C3gV2Z2PcHfra+4uwMfmdm/E4S9A0+7+/8dqoOR3s/S04kcgssAyeIdfQHo6EtA3HKw2MGXi/gvAu3t171p5x1s3/zu+2zf0ua47eO/nHSy75ay2DoOrG3NZc3H10G72u67eb/t77t1Wdz2Ih3IGD++297LWv0PHQLFxcVeXl6e7GaIiIh0m9j4reL2ynS/jYiISIgpqEVEREJMQS0iIhJiCmoREZEQU1CLiIiEmIJaREQkxBTUIiIiIaagFhERCTEFtYiISIgpqEVEREJMQS0iIhJiCmoREZEQU1CLiIiEmIJaREQkxBTUIiIiIaagFhERCTEFtYiISIgpqEVEREJMQS0iIhJiCmoREZEQU1CLiIiEmIJaREQkxBTUIiIiIaagFhERCbGEgtrM5pjZu2a2xszmt1NeYGbPmtnrZvaWmZ3TTvluM/t2VzVcRESkL9hvUJtZBFgAfAYYB1xqZuPaVLsZeMzdJwGXAL9oU/4z4E8H31wREZG+JZEe9WRgjbuvc/c64BHg/DZ1HBgQe50LbGoqMLPPA+uAdw6+uSIiIn1LIkE9HNgQt1wZWxfvVuALZlYJPA3MAzCzbOBfgO8fdEtFRET6oESC2tpZ522WLwUWuns+cA7wgJmlEAT0z9x9d6dvYHaVmZWbWfm2bdsSabeIiEifkJpAnUpgRNxyPnGntmOuBOYAuPtLZpYBDAZOAf7OzO4CBgJRM6tx95/Hb+zu9wL3AhQXF7f9EiAiItJnJRLUS4HRZlYIbCQYLHZZmzoVQAmw0MzGAhnANnc/vamCmd0K7G4b0iIiItKx/Z76dvcG4FrgGWAlwejud8zsNjM7L1btW8BXzexN4GHgK+6unrGIiMhBsrDlaXFxsfk7OwkAABWOSURBVJeXlye7GSIiIt3GzJa5e3F7ZXoymYiISIgpqEVEREJMQS0iIhJiCmoREZEQU1CLiIiEmIJaREQkxBTUIiIiIaagFhERCTEFtYiISIgpqEVEREJMQS0iIhJiCmoREZEQU1CLiIiEWK8P6p+//nOWbFhCbWNtspsiIiJywFKT3YBDaVftLn638ndU1VeRmZrJ6cNPp6SghNPzT6d/ev9kN09ERGS/enVQ5/bL5bmLn+PVLa9SWlFKWUUZi9YvIjUllVOOPIWSghJmjpjJ4MzByW6qiIhIu8zdk92GVoqLi728vPyQ7Lsx2shb29+idH0ppRWlVO6uxDAmHT6JWQWzKCkoIb9//iF5bxERkY6Y2TJ3L263rC8FdTx3Z/VHqymrKKO0opR3P3oXgGMPO5aSghJmFcxizGFjMLND3hYREenbFNQJ2FC1oTm039j6Bo4zov8ISgpKKCko4YQhJ5BivX7snYiIJIGC+gBtr97OsxuepbSilFc2v0JDtIHBmYOZOWImJQUlTD5iMmmRtKS2UUREeg8F9UGoqqvi+crnKa0o5YWNL1DdUE3/tP5MHzGdkoISTht2GllpWclupoiI9GAK6i5S01DDy5tfprSilCUblvBx7cf0i/Rj6rCplBSUcEb+GQzMGJjsZoqISA+joD4EGqINvL71dRavX0zZhjK27NlCxCIUDy1mVsEsZhXM4ojsI5LdTBER6QEU1IeYu7NixwpKK4LbvtbtWgfA+LzxlIwMRpAflXtUklspIiJhpaDuZu/vej8I7fWlLN+xHICjco9qHkE+Lm+cbvsSEZFmBx3UZjYHuBuIAL929x+2KS8A7gcGxurMd/enzWw28EMgHagD/tndyzp7r94Q1PG27NlCWUUZZRVllH9YTqM3ckT2EcwaETxg5cShJ5Ka0qsfECciIvtxUEFtZhFgNTAbqASWApe6+4q4OvcCr7v7L81sHPC0u48ys0nAh+6+yczGA8+4+/DO3q+3BXW8j2s+5rnK51hcsZiXNr1EbWMtA/sN5IwRZ1BSUMLUYVPpF+mX7GaKiEg36yyoE+nKTQbWuPu62M4eAc4HVsTVcWBA7HUusAnA3V+Pq/MOkGFm/dy9T05lNTBjIOcfcz7nH3M+e+v38uKmF5tPkT+55kkyUzOZNnwaJQUlTM+frolDREQkoaAeDmyIW64ETmlT51ZgkZnNA7KBM9vZz4UEve4+GdJtZaVlMXvkbGaPnE19Yz1LtyxlccVint3wLH9Z/xdNHCIiIkBip74vAs5293+MLX8RmOzu8+Lq3BDb10/NbCrwG2C8u0dj5ccDTwFnufvadt7jKuAqgIKCgpPWr1/fJQfXE0U9ylvb3moeQb6hagOGMfHwic3PIB/Rf0SymykiIl3oYK9RTwVudfezY8s3Arj7nXF13gHmuPuG2PI6YIq7bzWzfKAMuMLdX9xfY3vzNeoD5e689/F7zafHNXGIiEjvdLBBnUowmKwE2EgwmOwyd38nrs6fgEfdfaGZjQVKCU6Z5wLPAbe5+/9LpLEK6o41TRxSVlHG61tfx3Hyc/KD275GllA0pEgTh4iI9EBdcXvWOcB/ENx6dZ+732FmtwHl7v5UbKT3r4AcgoFl33H3RWZ2M3Aj8F7c7s5y960dvZeCOjHbq7ezZMMSFlcsbp44JC8jj5kFMzmz4ExNHCIi0oPogSe9XFVdFX+t/CulFaX8deNfmycOOT3/dEoKSpg2fJomDhERCTEFdR9S21jLy5teZnHF4tYThxw5lZKRmjhERCSMDvY+aulB+kX6MWPEDGaMmNE8cUjTCPIllUuIWISThp7ErILgyWiaOEREJNzUo+4j3J0VO1dQur71xCHH5x3PmSPP1MQhIiJJpFPfso+miUPKKsp4e/vbABTmFjZPHHJ83vG67UtEpJsoqKVTW/Zs4dkNz1JaUUr5lmDikKFZQ5lVMIszC87UxCEiIoeYgloS1jRxSGlFKX/b9LfmiUNm5M9onjgkIzUj2c0UEelVFNTyqeyt38vfNv2N0opSntvwHFX1VZo4RETkEFBQy0FrmjiktKKUsg1lbK/eHkwccsQpzCqYxayCWZo4RETkU1JQS5dqmjikrKKMxRWLmycOKRpS1DyCXBOHiIgkTkEth0z8xCFlFWWs2rkKgDGHjWkeQa6JQ0REOqeglm5TWVVJWUUZpRWlrSYOaXrAStGQIiIpkWQ3U0QkVBTUkhRNE4eUVpTy8uaXW00cUlJQwilHnKKJQ0REUFBLCOyu281fN8YmDqn8K3sb9pKTltM8ccjpw0/XxCEi0mcpqCVUmiYOKa0oZcmGJXxU+xHpKemcOuxUZhXM4owRZ3BYxmHJbqaISLfRpBwSKu1NHNJ0XXtJ5RJSLIWThp7UPBhNE4eISF+mHrWERvzEIWUVZazdtRYIJg5pCu2jBmriEBHpfXTqW3qkD3Z90DxFZ9PEIaMGjGoO7fGDx+u2LxHpFRTU0uN9uOdDyjaUtTtxSElBCScNPUkTh/Qx7k5DtIHqxmpqGmqoaaihuqGamsaaluW4sqbyfdbFLze21KuP1if7ECXEbj/tdk7PP73L9qegll5lV+2uYOKQ9aW8uOlFahtrye2X2zxxyKnDTtXEIUnWGG2ktrF23+CMW+6srCk0qxuq21+O1Wv0xgNuW3pKOhmpGWSkZpCZmklGJKNlOZLZ/Do9JV1nbKRDF425iGMHHdtl+1NQS6/VauKQyueoqmuZOGRWwSym509nQPqAZDczNNyd+mj9PoG3v9BMtBfatN+6aN0Bty3FUloF5z4h2mY5IxJb195yB2X9Iv30wB0JJQW19An10WDikLKKMsoqythWvY3UlFQmHzGZkoISZo6YyZCsIcluZocS7YXuN2TbhGrbnmjUowfctn6Rfu2GYXs90X2WOwnU5p+RTFJTUtWDlT5LQS19TvzEIaUVpVRUVTRPHNI0GG3EgMQmDnF36qJ1LaHYNiQ/bc+0zandQ9UL3V/PdH8hm5GaQYqlHHDbRCRxCmrp09ydNR+vaR5B3jRxyOjDRjPmsDHUNtS2O+goPkidA///pMPTtAcRmvE9U/VCRXoPBbVInKaJQ8o2lLFlz5aWXmdciB5wLzQ1s9W6fpF+6oWKSMIOOqjNbA5wNxABfu3uP2xTXgDcDwyM1Znv7k/Hym4ErgQagevc/ZnO3ktBLSIifc1BPULUzCLAAmA2UAksNbOn3H1FXLWbgcfc/ZdmNg54GhgVe30JcDwwDFhsZmPcP8U9FSIiIn1QIufmJgNr3H2du9cBjwDnt6njQNM9MLnAptjr84FH3L3W3d8H1sT2JyIiIglIJKiHAxvilitj6+LdCnzBzCoJetPzDmBbERER6UAiQd3ekNK2F7YvBRa6ez5wDvCAmaUkuC1mdpWZlZtZ+bZt2xJokoiISN+QSFBXAvE3nObTcmq7yZXAYwDu/hKQAQxOcFvc/V53L3b34iFDwvtAChERke6WSFAvBUabWaGZpRMMDnuqTZ0KoATAzMYSBPW2WL1LzKyfmRUCo4FXu6rxIiIivd1+R327e4OZXQs8Q3Dr1X3u/o6Z3QaUu/tTwLeAX5nZ9QSntr/iwX1f75jZY8AKoAH4ukZ8i4iIJE4PPBEREUmyzu6j1qOTREREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEFNQiIiIhpqAWEREJMQW1iIhIiCmoRUREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEFNQiIiIhpqAWEREJsV4f1GGbb1tERORApCa7AYdSNOpM+sFfOGJABkcNyQ7+Dc6Jvc4hNzMt2U0UERHpVK8O6tqGKH9fnM+6bXtYtaWKRSs+pDHa0sMenJPeHNxHD2kJ8BGHZZIa6fUnG0REpAfo1UGdmR7hps+Oa16ua4hSsXMv67btZt32PcHPbXtYtOJDdu7Z0FwvLWIUDMriqFh4Hx3XCx+UnZ6MQxERkT6qVwd1W+mpKRxzeA7HHJ6zT9nHe+tYu23PPiH+3LvbqGuMNtcbmJXGUYOzm0P8qME5HD0km4K8LPqlRrrzcEREpA/oU0HdmYFZ6Zw0Mp2TRh7Wan1DY5SNH1ezbtse1saF+HOrt/H4ssrmeikGIwZltRviQ/r3w8y6+5BERKQXUFDvR2okhZF52YzMy2bmcYe3KquqqWfdtj2s2x70vpvC/G9rd1Db0NIL798vlcIh2fuEeOHgbDLT1QsXEZGOKagPQv+MNIpGDKRoxMBW66NRZ9Ou6lh4N/XC9/Dq+zt58o1NreoOH5gZC+6WED96SA5HDMggJUW9cBGRvk5BfQikpBj5h2WRf1gW08cMaVW2t66B92PBHd8bf3xZJXvqGpvrZaZFKByc3TyI7eimXviQbHL66T+biEhfob/43SwrPZXjh+Vy/LDcVuvdna1VtcF18LgQf6tyF0+/vZm4u8oYOqBfq/vBm0amDz8sk4h64SIivUpCQW1mc4C7gQjwa3f/YZvynwEzY4tZwOHuPjBWdhfwWYKnoP0F+IbrcWH7MDOGDshg6IAMTj16cKuymvrG5tvK1saF+B/f2syu6vrmeumpKYzKy2o3xHOz9HAXEZGeaL9BbWYRYAEwG6gElprZU+6+oqmOu18fV38eMCn2+lTgNOCEWPELwAxgSRe1v0/ISIswZmh/xgzt32q9u7NzT12r28nWbtvD6q1VLF75IQ1x3fC87PR9nsx21JBsCgZlkaaHu4iIhFYiPerJwBp3XwdgZo8A5wMrOqh/KfC92GsHMoB0wIA04MODabC0MDPycvqRl9OPk0cNalVW39j0cJeWEF+3fTeLV37IjvK65nqpKU0Pd4mFd9ygtrzsdN1WJiKSZIkE9XBgQ9xyJXBKexXNbCRQCJQBuPtLZvYssJkgqH/u7isPqsWSkLRICkcPyeHoITnA0FZlu/bWs7b5lrKWEH9+9fZWD3cZkJHaaiR6U4iPzMsiI023lYmIdIdEgrq9LlVH15gvAR5390YAMzsGGAvkx8r/YmbT3f35Vm9gdhVwFUBBQUEi7ZaDkJuVxokFh3FiQeuHuzRGnY0fVe8T4i+u2c4fXtvYXC/FIP+wrDan0oMwP1wPdxER6VKJBHUlMCJuOR/Y1EHdS4Cvxy3PBV52990AZvYnYArQKqjd/V7gXoDi4mINNEuSSIpRkJdFQV4WM49tXba7toH3Yz3vtXEh/sq6nVTXt9xWltMvteW2srgQP2pwjh7uIiLyKSQS1EuB0WZWCGwkCOPL2lYys2OBw4CX4lZXAF81szsJeuYzgP842EZL98vpl8qE/Fwm5Le+rSwadbZ8UtPqnvC123ZT/sFHPPXmJuLH9w/LzYh7MlvLtfBhuZl6uIuISAf2G9Tu3mBm1wLPENyedZ+7v2NmtwHl7v5UrOqlwCNtbr16HJgFvE1wuvzP7v6/XXoEklQpKcawgZkMG5jJtNH73lbW8nCXluekP/HaRqpqG5rrZaSlMCovfqrRlt54/wzdViYifZuF7Zbm4uJiLy8vT3Yz5BByd7ZV1Qan0OOvh2/fw4ade1s93GVI/37Nve+j40I8X3OGi0gvYmbL3L24vTI9mUy6nZlx+IAMDh+QwdSj81qV1TY0UrFj7z4h/qflm/l4b9zDXSIpFORlxcI7p1WYD8zSnOEi0nsoqCVU+qVGGD20P6PbPNwFCB7u0vRgl1iIr9m6m7JVW6lvbOmGD8pOjwV36xAvGJRFeqp64YeCu+MOUXeizT9bXnu0ZV1j27rRYLkxVu7x+4i2s6+47fbZV1N5lFhZfHuCdjZG29lX3D7j64p0ZPa4oYzMy+6W91JQS48xKDudQdmDKG7zcJeGxigbPqpudU/42m17KFu1jcfKW+YMjzQ93CUuxAdlp7f7xzzqTmO0/T/m8X/svU2IRKPt76ttCETjAqajfbkTK4sPo8731V64xbe/Mdq6bnMbYtu1e6yx4+o0iJVp0scUDMpSUIskKjWSQuHgbAoHZ1MytnXZrur62IC23a1Gpr+wZnurOcMPlRSDFLPgX0rLa2teH/uZYq3qNpVHUvata7HXQVn8dsFlhZQUSEtJaVW31XvF6ljT6/j1ze2hzbZx5XFtisS/b5vySHvHmmL7vJfFvWfb94pvQ+vPoqU8krJv3ab9RlIObF/xn6seByCdyezGhz4pqKVXy81MY+KIgUxsZ87wjR9Xs6u6/qD+mHcUrk11RUQOloJa+qSUFGPEoKxWT/IREQkjjawREREJMQW1iIhIiCmoRUREQkxBLSIiEmIKahERkRBTUIuIiISYglpERCTEFNQiIiIhpqAWEREJMQW1iIhIiFnYpnIzs23A+i7e7WBgexfvU3oH/W5IR/S7IZ3p6t+Pke4+pL2C0AX1oWBm5e5enOx2SPjod0M6ot8N6Ux3/n7o1LeIiEiIKahFRERCrK8E9b3JboCEln43pCP63ZDOdNvvR5+4Ri0iItJT9ZUetYiISI/Uo4PazEaZ2fI2684wMzezz8Wt+6OZnRF7vcTMyuPKis1sSXe1WUTCzcx2J7sN0nOZ2QdmNrgr99mjg7oTlcBNnZQfbmaf6a7GiEjPZmaRZLdB+q5eE9RmdpSZvQ6cDLwJ7DKz2R1U/zFwc7c1TkIhdgZmpZn9yszeMbNFZpYZO8vyIzN71cxWm9npyW6rJF/s7NyzZvY74O1kt0e6j5llm9n/mdmbZrbczC42s3PMbJWZvWBm95jZH2N182J/S143s/8CrKvb0yuC2syOBf4fcAWwNLb6djoO45eAWjOb2Q3Nk3AZDSxw9+OBj4ELY+tT3X0y8E3ge8lqnITOZOAmdx+X7IZIt5oDbHL3IncfD/wZ+C/gM+4+DYh/gtj3gBfcfRLwFFDQ1Y3pDUE9BPgf4Avu/kbTSnf/K0AnvaPOglx6r/fjfk+WAaNir//QzjqRV939/WQ3Qrrd28CZsTNtpwOFwLq434WH4+pOBx4EcPf/Az7q6sb0hqDeBWwATmun7A46uFbt7mVABjDl0DVNQqg27nUjkNpmffw6kT3JboB0P3dfDZxEENh3Aufvb5ND2Z7eENR1wOeBL5nZZfEF7r4IOAwo6mDbO4DvHNrmiYhIT2Jmw4C97v4g8BPgVOAoMxsVq3JxXPXngctj232GIHO6VK/oObj7HjM7F/gLwSnteHcQnBpvb7unY7N1iYiINJkA/NjMokA98DXgSODPZrYdeDWu7veBh83sNeA5oKKrG6Mnk4mIiOyHmeW4+24zM2AB8J67/6w73rs3nPoWERE51L5qZm8A7wC5BKPAu4V61CIiIiGmHrWIiEiIKahFRERCTEEtIiISYgpqERGREFNQi4iIhJiCWkREJMT+PwYbch1Fs7wNAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TP   FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7  kNN  kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6   nn   nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "2                k5   nn   nn  kNN  kNN      kNN     nn       kNN      kNN"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k4'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>nn</td>\n",
       "      <td>5980</td>\n",
       "      <td>1427</td>\n",
       "      <td>871</td>\n",
       "      <td>581</td>\n",
       "      <td>0.773338</td>\n",
       "      <td>0.911446</td>\n",
       "      <td>0.807344</td>\n",
       "      <td>0.856243</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP   TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "12         nn  5980  1427  871  581  0.773338  0.911446   0.807344  0.856243"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TP   FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7  kNN  kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6   nn   nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "2                k5   nn   nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "3                k4   nn   nn   nn   nn       nn     nn        nn       nn"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k3'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>dtree</td>\n",
       "      <td>6096</td>\n",
       "      <td>1559</td>\n",
       "      <td>739</td>\n",
       "      <td>465</td>\n",
       "      <td>0.771532</td>\n",
       "      <td>0.929127</td>\n",
       "      <td>0.796342</td>\n",
       "      <td>0.857625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>SVM</td>\n",
       "      <td>6096</td>\n",
       "      <td>1559</td>\n",
       "      <td>739</td>\n",
       "      <td>465</td>\n",
       "      <td>0.771532</td>\n",
       "      <td>0.929127</td>\n",
       "      <td>0.796342</td>\n",
       "      <td>0.857625</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>nn</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>lr</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>sgd</td>\n",
       "      <td>6006</td>\n",
       "      <td>1459</td>\n",
       "      <td>839</td>\n",
       "      <td>555</td>\n",
       "      <td>0.772661</td>\n",
       "      <td>0.915409</td>\n",
       "      <td>0.804555</td>\n",
       "      <td>0.856410</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP   TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "5       dtree  6096  1559  739  465  0.771532  0.929127   0.796342  0.857625\n",
       "8         SVM  6096  1559  739  465  0.771532  0.929127   0.796342  0.857625\n",
       "13         nn  6006  1459  839  555  0.772661  0.915409   0.804555  0.856410\n",
       "16         lr  6006  1459  839  555  0.772661  0.915409   0.804555  0.856410\n",
       "19        sgd  6006  1459  839  555  0.772661  0.915409   0.804555  0.856410"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k3</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TP   FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7  kNN  kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6   nn   nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "2                k5   nn   nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "3                k4   nn   nn   nn   nn       nn     nn        nn       nn\n",
       "4                k3  SVM  SVM  sgd  sgd      sgd    SVM       sgd      SVM"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k2'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>dtree</td>\n",
       "      <td>6311</td>\n",
       "      <td>1789</td>\n",
       "      <td>509</td>\n",
       "      <td>250</td>\n",
       "      <td>0.769839</td>\n",
       "      <td>0.961896</td>\n",
       "      <td>0.779136</td>\n",
       "      <td>0.860924</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>SVM</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>nn</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>lr</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>sgd</td>\n",
       "      <td>6081</td>\n",
       "      <td>1557</td>\n",
       "      <td>741</td>\n",
       "      <td>480</td>\n",
       "      <td>0.770064</td>\n",
       "      <td>0.926840</td>\n",
       "      <td>0.796151</td>\n",
       "      <td>0.856539</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP   TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "6       dtree  6311  1789  509  250  0.769839  0.961896   0.779136  0.860924\n",
       "9         SVM  6081  1557  741  480  0.770064  0.926840   0.796151  0.856539\n",
       "14         nn  6081  1557  741  480  0.770064  0.926840   0.796151  0.856539\n",
       "17         lr  6081  1557  741  480  0.770064  0.926840   0.796151  0.856539\n",
       "20        sgd  6081  1557  741  480  0.770064  0.926840   0.796151  0.856539"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k3</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>k2</td>\n",
       "      <td>dtree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TP     FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7    kNN    kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6     nn     nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "2                k5     nn     nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "3                k4     nn     nn   nn   nn       nn     nn        nn       nn\n",
       "4                k3    SVM    SVM  sgd  sgd      sgd    SVM       sgd      SVM\n",
       "5                k2  dtree  dtree  sgd  sgd      sgd  dtree       sgd    dtree"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #Tree-based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='Tree'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kNN</td>\n",
       "      <td>5904</td>\n",
       "      <td>1188</td>\n",
       "      <td>1110</td>\n",
       "      <td>657</td>\n",
       "      <td>0.791737</td>\n",
       "      <td>0.899863</td>\n",
       "      <td>0.832487</td>\n",
       "      <td>0.864865</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>gnb</td>\n",
       "      <td>5979</td>\n",
       "      <td>1514</td>\n",
       "      <td>784</td>\n",
       "      <td>582</td>\n",
       "      <td>0.763404</td>\n",
       "      <td>0.911294</td>\n",
       "      <td>0.797945</td>\n",
       "      <td>0.850861</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>dtree</td>\n",
       "      <td>6257</td>\n",
       "      <td>1686</td>\n",
       "      <td>612</td>\n",
       "      <td>304</td>\n",
       "      <td>0.775370</td>\n",
       "      <td>0.953666</td>\n",
       "      <td>0.787738</td>\n",
       "      <td>0.862796</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Classifier    TP    FP    TN   FN  Accuracy    Recall  Precision  F1_Score\n",
       "3        kNN  5904  1188  1110  657  0.791737  0.899863   0.832487  0.864865\n",
       "4        gnb  5979  1514   784  582  0.763404  0.911294   0.797945  0.850861\n",
       "7      dtree  6257  1686   612  304  0.775370  0.953666   0.787738  0.862796"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k3</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>k2</td>\n",
       "      <td>dtree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Tree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TP     FP   TN   FN Accuracy Recall Precision F1_Score\n",
       "0                k7    kNN    kNN  kNN  kNN      kNN    kNN       kNN      kNN\n",
       "1                k6     nn     nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "2                k5     nn     nn  kNN  kNN      kNN     nn       kNN      kNN\n",
       "3                k4     nn     nn   nn   nn       nn     nn        nn       nn\n",
       "4                k3    SVM    SVM  sgd  sgd      sgd    SVM       sgd      SVM\n",
       "5                k2  dtree  dtree  sgd  sgd      sgd  dtree       sgd    dtree\n",
       "6              Tree  dtree  dtree  kNN  kNN      kNN  dtree       kNN      kNN"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k7</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k5</td>\n",
       "      <td>kNN</td>\n",
       "      <td>nn</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "      <td>nn</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k3</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>k2</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>dtree</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Tree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection Accuracy Recall Precision F1_Score\n",
       "0                k7      kNN    kNN       kNN      kNN\n",
       "1                k6      kNN     nn       kNN      kNN\n",
       "2                k5      kNN     nn       kNN      kNN\n",
       "3                k4       nn     nn        nn       nn\n",
       "4                k3      sgd    SVM       sgd      SVM\n",
       "5                k2      sgd  dtree       sgd    dtree\n",
       "6              Tree      kNN  dtree       kNN      kNN"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result[['Feature Selection', 'Accuracy', 'Recall', 'Precision', 'F1_Score']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Russia\n",
    "Country='Russia'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "data=pd.read_csv('/Volumes/BECCA/Project/Data/Result/%s/By_model/Total_Result.csv' %Country, index_col=[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4309</td>\n",
       "      <td>1038</td>\n",
       "      <td>4167</td>\n",
       "      <td>1319</td>\n",
       "      <td>0.782424</td>\n",
       "      <td>0.765636</td>\n",
       "      <td>0.805872</td>\n",
       "      <td>0.785239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4297</td>\n",
       "      <td>1151</td>\n",
       "      <td>4054</td>\n",
       "      <td>1331</td>\n",
       "      <td>0.770885</td>\n",
       "      <td>0.763504</td>\n",
       "      <td>0.788730</td>\n",
       "      <td>0.775912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4284</td>\n",
       "      <td>1155</td>\n",
       "      <td>4050</td>\n",
       "      <td>1344</td>\n",
       "      <td>0.769316</td>\n",
       "      <td>0.761194</td>\n",
       "      <td>0.787645</td>\n",
       "      <td>0.774194</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>L1</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4321</td>\n",
       "      <td>1051</td>\n",
       "      <td>4154</td>\n",
       "      <td>1307</td>\n",
       "      <td>0.782332</td>\n",
       "      <td>0.767768</td>\n",
       "      <td>0.804356</td>\n",
       "      <td>0.785636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Tree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4426</td>\n",
       "      <td>1278</td>\n",
       "      <td>3927</td>\n",
       "      <td>1202</td>\n",
       "      <td>0.771070</td>\n",
       "      <td>0.786425</td>\n",
       "      <td>0.775947</td>\n",
       "      <td>0.781151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>k4</td>\n",
       "      <td>gnb</td>\n",
       "      <td>3888</td>\n",
       "      <td>1974</td>\n",
       "      <td>3231</td>\n",
       "      <td>1740</td>\n",
       "      <td>0.657159</td>\n",
       "      <td>0.690832</td>\n",
       "      <td>0.663255</td>\n",
       "      <td>0.676762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>L1</td>\n",
       "      <td>gnb</td>\n",
       "      <td>4133</td>\n",
       "      <td>2120</td>\n",
       "      <td>3085</td>\n",
       "      <td>1495</td>\n",
       "      <td>0.666297</td>\n",
       "      <td>0.734364</td>\n",
       "      <td>0.660963</td>\n",
       "      <td>0.695733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>Tree</td>\n",
       "      <td>gnb</td>\n",
       "      <td>3897</td>\n",
       "      <td>2001</td>\n",
       "      <td>3204</td>\n",
       "      <td>1731</td>\n",
       "      <td>0.655497</td>\n",
       "      <td>0.692431</td>\n",
       "      <td>0.660732</td>\n",
       "      <td>0.676210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>k6</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3556</td>\n",
       "      <td>1211</td>\n",
       "      <td>3994</td>\n",
       "      <td>2072</td>\n",
       "      <td>0.696945</td>\n",
       "      <td>0.631841</td>\n",
       "      <td>0.745962</td>\n",
       "      <td>0.684175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>L1</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3556</td>\n",
       "      <td>1211</td>\n",
       "      <td>3994</td>\n",
       "      <td>2072</td>\n",
       "      <td>0.696945</td>\n",
       "      <td>0.631841</td>\n",
       "      <td>0.745962</td>\n",
       "      <td>0.684175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>Tree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>k6</td>\n",
       "      <td>SVM</td>\n",
       "      <td>3792</td>\n",
       "      <td>1549</td>\n",
       "      <td>3656</td>\n",
       "      <td>1836</td>\n",
       "      <td>0.687529</td>\n",
       "      <td>0.673774</td>\n",
       "      <td>0.709979</td>\n",
       "      <td>0.691403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>k5</td>\n",
       "      <td>SVM</td>\n",
       "      <td>4485</td>\n",
       "      <td>2676</td>\n",
       "      <td>2529</td>\n",
       "      <td>1143</td>\n",
       "      <td>0.647466</td>\n",
       "      <td>0.796908</td>\n",
       "      <td>0.626309</td>\n",
       "      <td>0.701384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>k4</td>\n",
       "      <td>SVM</td>\n",
       "      <td>4487</td>\n",
       "      <td>2676</td>\n",
       "      <td>2529</td>\n",
       "      <td>1141</td>\n",
       "      <td>0.647651</td>\n",
       "      <td>0.797264</td>\n",
       "      <td>0.626414</td>\n",
       "      <td>0.701587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>L1</td>\n",
       "      <td>SVM</td>\n",
       "      <td>3787</td>\n",
       "      <td>1545</td>\n",
       "      <td>3660</td>\n",
       "      <td>1841</td>\n",
       "      <td>0.687437</td>\n",
       "      <td>0.672886</td>\n",
       "      <td>0.710240</td>\n",
       "      <td>0.691058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>Tree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>4581</td>\n",
       "      <td>2868</td>\n",
       "      <td>2337</td>\n",
       "      <td>1047</td>\n",
       "      <td>0.638604</td>\n",
       "      <td>0.813966</td>\n",
       "      <td>0.614982</td>\n",
       "      <td>0.700619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>k6</td>\n",
       "      <td>nn</td>\n",
       "      <td>4233</td>\n",
       "      <td>2657</td>\n",
       "      <td>2548</td>\n",
       "      <td>1395</td>\n",
       "      <td>0.625958</td>\n",
       "      <td>0.752132</td>\n",
       "      <td>0.614369</td>\n",
       "      <td>0.676306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>k5</td>\n",
       "      <td>nn</td>\n",
       "      <td>4070</td>\n",
       "      <td>2742</td>\n",
       "      <td>2463</td>\n",
       "      <td>1558</td>\n",
       "      <td>0.603065</td>\n",
       "      <td>0.723170</td>\n",
       "      <td>0.597475</td>\n",
       "      <td>0.654341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>k4</td>\n",
       "      <td>nn</td>\n",
       "      <td>3849</td>\n",
       "      <td>2465</td>\n",
       "      <td>2740</td>\n",
       "      <td>1779</td>\n",
       "      <td>0.608234</td>\n",
       "      <td>0.683902</td>\n",
       "      <td>0.609598</td>\n",
       "      <td>0.644616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>L1</td>\n",
       "      <td>nn</td>\n",
       "      <td>4227</td>\n",
       "      <td>2618</td>\n",
       "      <td>2587</td>\n",
       "      <td>1401</td>\n",
       "      <td>0.629004</td>\n",
       "      <td>0.751066</td>\n",
       "      <td>0.617531</td>\n",
       "      <td>0.677784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>Tree</td>\n",
       "      <td>nn</td>\n",
       "      <td>3734</td>\n",
       "      <td>2260</td>\n",
       "      <td>2945</td>\n",
       "      <td>1894</td>\n",
       "      <td>0.616542</td>\n",
       "      <td>0.663468</td>\n",
       "      <td>0.622956</td>\n",
       "      <td>0.642574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>k6</td>\n",
       "      <td>lr</td>\n",
       "      <td>4171</td>\n",
       "      <td>2571</td>\n",
       "      <td>2634</td>\n",
       "      <td>1457</td>\n",
       "      <td>0.628173</td>\n",
       "      <td>0.741116</td>\n",
       "      <td>0.618659</td>\n",
       "      <td>0.674373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>k5</td>\n",
       "      <td>lr</td>\n",
       "      <td>3862</td>\n",
       "      <td>2471</td>\n",
       "      <td>2734</td>\n",
       "      <td>1766</td>\n",
       "      <td>0.608880</td>\n",
       "      <td>0.686212</td>\n",
       "      <td>0.609822</td>\n",
       "      <td>0.645765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>k4</td>\n",
       "      <td>lr</td>\n",
       "      <td>3865</td>\n",
       "      <td>2474</td>\n",
       "      <td>2731</td>\n",
       "      <td>1763</td>\n",
       "      <td>0.608880</td>\n",
       "      <td>0.686745</td>\n",
       "      <td>0.609718</td>\n",
       "      <td>0.645943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>L1</td>\n",
       "      <td>lr</td>\n",
       "      <td>4172</td>\n",
       "      <td>2574</td>\n",
       "      <td>2631</td>\n",
       "      <td>1456</td>\n",
       "      <td>0.627989</td>\n",
       "      <td>0.741294</td>\n",
       "      <td>0.618441</td>\n",
       "      <td>0.674317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>Tree</td>\n",
       "      <td>lr</td>\n",
       "      <td>3870</td>\n",
       "      <td>2458</td>\n",
       "      <td>2747</td>\n",
       "      <td>1758</td>\n",
       "      <td>0.610819</td>\n",
       "      <td>0.687633</td>\n",
       "      <td>0.611568</td>\n",
       "      <td>0.647374</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>k6</td>\n",
       "      <td>sgd</td>\n",
       "      <td>4735</td>\n",
       "      <td>3499</td>\n",
       "      <td>1706</td>\n",
       "      <td>893</td>\n",
       "      <td>0.594572</td>\n",
       "      <td>0.841329</td>\n",
       "      <td>0.575055</td>\n",
       "      <td>0.683163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>k5</td>\n",
       "      <td>sgd</td>\n",
       "      <td>3561</td>\n",
       "      <td>2248</td>\n",
       "      <td>2957</td>\n",
       "      <td>2067</td>\n",
       "      <td>0.601680</td>\n",
       "      <td>0.632729</td>\n",
       "      <td>0.613014</td>\n",
       "      <td>0.622716</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>k4</td>\n",
       "      <td>sgd</td>\n",
       "      <td>3679</td>\n",
       "      <td>2359</td>\n",
       "      <td>2846</td>\n",
       "      <td>1949</td>\n",
       "      <td>0.602326</td>\n",
       "      <td>0.653696</td>\n",
       "      <td>0.609308</td>\n",
       "      <td>0.630722</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>L1</td>\n",
       "      <td>sgd</td>\n",
       "      <td>4567</td>\n",
       "      <td>3123</td>\n",
       "      <td>2082</td>\n",
       "      <td>1061</td>\n",
       "      <td>0.613773</td>\n",
       "      <td>0.811478</td>\n",
       "      <td>0.593888</td>\n",
       "      <td>0.685839</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>Tree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>4048</td>\n",
       "      <td>2980</td>\n",
       "      <td>2225</td>\n",
       "      <td>1580</td>\n",
       "      <td>0.579064</td>\n",
       "      <td>0.719261</td>\n",
       "      <td>0.575982</td>\n",
       "      <td>0.639697</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Model Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  \\\n",
       "0     k6        kNN  4309  1038  4167  1319  0.782424  0.765636   0.805872   \n",
       "1     k5        kNN  4297  1151  4054  1331  0.770885  0.763504   0.788730   \n",
       "2     k4        kNN  4284  1155  4050  1344  0.769316  0.761194   0.787645   \n",
       "3     L1        kNN  4321  1051  4154  1307  0.782332  0.767768   0.804356   \n",
       "4   Tree        kNN  4426  1278  3927  1202  0.771070  0.786425   0.775947   \n",
       "5     k4        gnb  3888  1974  3231  1740  0.657159  0.690832   0.663255   \n",
       "6     L1        gnb  4133  2120  3085  1495  0.666297  0.734364   0.660963   \n",
       "7   Tree        gnb  3897  2001  3204  1731  0.655497  0.692431   0.660732   \n",
       "8     k6      dtree  3556  1211  3994  2072  0.696945  0.631841   0.745962   \n",
       "9     k5      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868   \n",
       "10    k4      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868   \n",
       "11    L1      dtree  3556  1211  3994  2072  0.696945  0.631841   0.745962   \n",
       "12  Tree      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868   \n",
       "13    k6        SVM  3792  1549  3656  1836  0.687529  0.673774   0.709979   \n",
       "14    k5        SVM  4485  2676  2529  1143  0.647466  0.796908   0.626309   \n",
       "15    k4        SVM  4487  2676  2529  1141  0.647651  0.797264   0.626414   \n",
       "16    L1        SVM  3787  1545  3660  1841  0.687437  0.672886   0.710240   \n",
       "17  Tree        SVM  4581  2868  2337  1047  0.638604  0.813966   0.614982   \n",
       "18    k6         nn  4233  2657  2548  1395  0.625958  0.752132   0.614369   \n",
       "19    k5         nn  4070  2742  2463  1558  0.603065  0.723170   0.597475   \n",
       "20    k4         nn  3849  2465  2740  1779  0.608234  0.683902   0.609598   \n",
       "21    L1         nn  4227  2618  2587  1401  0.629004  0.751066   0.617531   \n",
       "22  Tree         nn  3734  2260  2945  1894  0.616542  0.663468   0.622956   \n",
       "23    k6         lr  4171  2571  2634  1457  0.628173  0.741116   0.618659   \n",
       "24    k5         lr  3862  2471  2734  1766  0.608880  0.686212   0.609822   \n",
       "25    k4         lr  3865  2474  2731  1763  0.608880  0.686745   0.609718   \n",
       "26    L1         lr  4172  2574  2631  1456  0.627989  0.741294   0.618441   \n",
       "27  Tree         lr  3870  2458  2747  1758  0.610819  0.687633   0.611568   \n",
       "28    k6        sgd  4735  3499  1706   893  0.594572  0.841329   0.575055   \n",
       "29    k5        sgd  3561  2248  2957  2067  0.601680  0.632729   0.613014   \n",
       "30    k4        sgd  3679  2359  2846  1949  0.602326  0.653696   0.609308   \n",
       "31    L1        sgd  4567  3123  2082  1061  0.613773  0.811478   0.593888   \n",
       "32  Tree        sgd  4048  2980  2225  1580  0.579064  0.719261   0.575982   \n",
       "\n",
       "    F1_Score  \n",
       "0   0.785239  \n",
       "1   0.775912  \n",
       "2   0.774194  \n",
       "3   0.785636  \n",
       "4   0.781151  \n",
       "5   0.676762  \n",
       "6   0.695733  \n",
       "7   0.676210  \n",
       "8   0.684175  \n",
       "9   0.678095  \n",
       "10  0.678095  \n",
       "11  0.684175  \n",
       "12  0.678095  \n",
       "13  0.691403  \n",
       "14  0.701384  \n",
       "15  0.701587  \n",
       "16  0.691058  \n",
       "17  0.700619  \n",
       "18  0.676306  \n",
       "19  0.654341  \n",
       "20  0.644616  \n",
       "21  0.677784  \n",
       "22  0.642574  \n",
       "23  0.674373  \n",
       "24  0.645765  \n",
       "25  0.645943  \n",
       "26  0.674317  \n",
       "27  0.647374  \n",
       "28  0.683163  \n",
       "29  0.622716  \n",
       "30  0.630722  \n",
       "31  0.685839  \n",
       "32  0.639697  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# By feature"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #All"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k6'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kNN</td>\n",
       "      <td>4309</td>\n",
       "      <td>1038</td>\n",
       "      <td>4167</td>\n",
       "      <td>1319</td>\n",
       "      <td>0.782424</td>\n",
       "      <td>0.765636</td>\n",
       "      <td>0.805872</td>\n",
       "      <td>0.785239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>dtree</td>\n",
       "      <td>3556</td>\n",
       "      <td>1211</td>\n",
       "      <td>3994</td>\n",
       "      <td>2072</td>\n",
       "      <td>0.696945</td>\n",
       "      <td>0.631841</td>\n",
       "      <td>0.745962</td>\n",
       "      <td>0.684175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>SVM</td>\n",
       "      <td>3792</td>\n",
       "      <td>1549</td>\n",
       "      <td>3656</td>\n",
       "      <td>1836</td>\n",
       "      <td>0.687529</td>\n",
       "      <td>0.673774</td>\n",
       "      <td>0.709979</td>\n",
       "      <td>0.691403</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>nn</td>\n",
       "      <td>4233</td>\n",
       "      <td>2657</td>\n",
       "      <td>2548</td>\n",
       "      <td>1395</td>\n",
       "      <td>0.625958</td>\n",
       "      <td>0.752132</td>\n",
       "      <td>0.614369</td>\n",
       "      <td>0.676306</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>lr</td>\n",
       "      <td>4171</td>\n",
       "      <td>2571</td>\n",
       "      <td>2634</td>\n",
       "      <td>1457</td>\n",
       "      <td>0.628173</td>\n",
       "      <td>0.741116</td>\n",
       "      <td>0.618659</td>\n",
       "      <td>0.674373</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>sgd</td>\n",
       "      <td>4735</td>\n",
       "      <td>3499</td>\n",
       "      <td>1706</td>\n",
       "      <td>893</td>\n",
       "      <td>0.594572</td>\n",
       "      <td>0.841329</td>\n",
       "      <td>0.575055</td>\n",
       "      <td>0.683163</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  F1_Score\n",
       "0         kNN  4309  1038  4167  1319  0.782424  0.765636   0.805872  0.785239\n",
       "8       dtree  3556  1211  3994  2072  0.696945  0.631841   0.745962  0.684175\n",
       "13        SVM  3792  1549  3656  1836  0.687529  0.673774   0.709979  0.691403\n",
       "18         nn  4233  2657  2548  1395  0.625958  0.752132   0.614369  0.676306\n",
       "23         lr  4171  2571  2634  1457  0.628173  0.741116   0.618659  0.674373\n",
       "28        sgd  4735  3499  1706   893  0.594572  0.841329   0.575055  0.683163"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection   TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6  kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=pd.DataFrame(opt, index=[0])\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k5'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>kNN</td>\n",
       "      <td>4297</td>\n",
       "      <td>1151</td>\n",
       "      <td>4054</td>\n",
       "      <td>1331</td>\n",
       "      <td>0.770885</td>\n",
       "      <td>0.763504</td>\n",
       "      <td>0.788730</td>\n",
       "      <td>0.775912</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>SVM</td>\n",
       "      <td>4485</td>\n",
       "      <td>2676</td>\n",
       "      <td>2529</td>\n",
       "      <td>1143</td>\n",
       "      <td>0.647466</td>\n",
       "      <td>0.796908</td>\n",
       "      <td>0.626309</td>\n",
       "      <td>0.701384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>nn</td>\n",
       "      <td>4070</td>\n",
       "      <td>2742</td>\n",
       "      <td>2463</td>\n",
       "      <td>1558</td>\n",
       "      <td>0.603065</td>\n",
       "      <td>0.723170</td>\n",
       "      <td>0.597475</td>\n",
       "      <td>0.654341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>lr</td>\n",
       "      <td>3862</td>\n",
       "      <td>2471</td>\n",
       "      <td>2734</td>\n",
       "      <td>1766</td>\n",
       "      <td>0.608880</td>\n",
       "      <td>0.686212</td>\n",
       "      <td>0.609822</td>\n",
       "      <td>0.645765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>sgd</td>\n",
       "      <td>3561</td>\n",
       "      <td>2248</td>\n",
       "      <td>2957</td>\n",
       "      <td>2067</td>\n",
       "      <td>0.601680</td>\n",
       "      <td>0.632729</td>\n",
       "      <td>0.613014</td>\n",
       "      <td>0.622716</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  F1_Score\n",
       "1         kNN  4297  1151  4054  1331  0.770885  0.763504   0.788730  0.775912\n",
       "9       dtree  3303   811  4394  2325  0.710514  0.586887   0.802868  0.678095\n",
       "14        SVM  4485  2676  2529  1143  0.647466  0.796908   0.626309  0.701384\n",
       "19         nn  4070  2742  2463  1558  0.603065  0.723170   0.597475  0.654341\n",
       "24         lr  3862  2471  2734  1766  0.608880  0.686212   0.609822  0.645765\n",
       "29        sgd  3561  2248  2957  2067  0.601680  0.632729   0.613014  0.622716"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k4'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>kNN</td>\n",
       "      <td>4284</td>\n",
       "      <td>1155</td>\n",
       "      <td>4050</td>\n",
       "      <td>1344</td>\n",
       "      <td>0.769316</td>\n",
       "      <td>0.761194</td>\n",
       "      <td>0.787645</td>\n",
       "      <td>0.774194</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>gnb</td>\n",
       "      <td>3888</td>\n",
       "      <td>1974</td>\n",
       "      <td>3231</td>\n",
       "      <td>1740</td>\n",
       "      <td>0.657159</td>\n",
       "      <td>0.690832</td>\n",
       "      <td>0.663255</td>\n",
       "      <td>0.676762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>SVM</td>\n",
       "      <td>4487</td>\n",
       "      <td>2676</td>\n",
       "      <td>2529</td>\n",
       "      <td>1141</td>\n",
       "      <td>0.647651</td>\n",
       "      <td>0.797264</td>\n",
       "      <td>0.626414</td>\n",
       "      <td>0.701587</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>nn</td>\n",
       "      <td>3849</td>\n",
       "      <td>2465</td>\n",
       "      <td>2740</td>\n",
       "      <td>1779</td>\n",
       "      <td>0.608234</td>\n",
       "      <td>0.683902</td>\n",
       "      <td>0.609598</td>\n",
       "      <td>0.644616</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>lr</td>\n",
       "      <td>3865</td>\n",
       "      <td>2474</td>\n",
       "      <td>2731</td>\n",
       "      <td>1763</td>\n",
       "      <td>0.608880</td>\n",
       "      <td>0.686745</td>\n",
       "      <td>0.609718</td>\n",
       "      <td>0.645943</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>sgd</td>\n",
       "      <td>3679</td>\n",
       "      <td>2359</td>\n",
       "      <td>2846</td>\n",
       "      <td>1949</td>\n",
       "      <td>0.602326</td>\n",
       "      <td>0.653696</td>\n",
       "      <td>0.609308</td>\n",
       "      <td>0.630722</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  F1_Score\n",
       "2         kNN  4284  1155  4050  1344  0.769316  0.761194   0.787645  0.774194\n",
       "5         gnb  3888  1974  3231  1740  0.657159  0.690832   0.663255  0.676762\n",
       "10      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868  0.678095\n",
       "15        SVM  4487  2676  2529  1141  0.647651  0.797264   0.626414  0.701587\n",
       "20         nn  3849  2465  2740  1779  0.608234  0.683902   0.609598  0.644616\n",
       "25         lr  3865  2474  2731  1763  0.608880  0.686745   0.609718  0.645943\n",
       "30        sgd  3679  2359  2846  1949  0.602326  0.653696   0.609308  0.630722"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM\n",
       "2                k4  dtree      kNN     dtree      kNN  dtree  SVM  SVM    SVM"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k3'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [Classifier, TP, FP, TN, FN, Accuracy, Recall, Precision, F1_Score]\n",
       "Index: []"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAewAAAEvCAYAAACHVvJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAfTklEQVR4nO3de3RW1Z3/8fd3AhQVocPFa7jNEquUEKAhojJSB2ixU6F4p1StQ2utA05rpzN0tNaf2pmpvUxlLabqqCtKLVBdldIZnTogjPWCgoqKIBotlhhFitVCkYth//7IYyaGBAI5JJzk/VorK+eyn7P3fhLyYZ9znrMjpYQkSTq4/VlbN0CSJO2dgS1JUg4Y2JIk5YCBLUlSDhjYkiTlgIEtSVIOdGrrBjSld+/eacCAAW3dDEmSWs1TTz31+5RSn8b2HbSBPWDAAFasWNHWzZAkqdVExGtN7cvklHhE3BERb0XEqib2T42I5wpfj0VEaRb1SpLUUWR1DbsCmLCH/b8FxqSUhgLXA7dmVK8kSR1CJqfEU0oPR8SAPex/rN7qMqA4i3olSeoo2uIa9jTggTaoV5KUgZ07d1JVVcW2bdvauim51bVrV4qLi+ncuXOzX9OqgR0Rp1Mb2KOb2H8pcClAv379WrFlkqTmqqqq4vDDD2fAgAFERFs3J3dSSmzatImqqioGDhzY7Ne12uewI2IocBswKaW0qbEyKaVbU0plKaWyPn0avatdktTGtm3bRq9evQzr/RQR9OrVa5/PULRKYEdEP+AXwIUppZdao05J0oFjWLfM/rx/WX2say7wOPCxiKiKiGkRcVlEXFYocg3QC/j3iFgZEX7AWpLUIvfddx8RwYsvvtjWTWkVWd0lPmUv+78EfCmLuiRJApg7dy6jR49m3rx5XHvttQekjpqaGoqKig7IsfeVzxKXJOXOli1bePTRR7n99tuZN29e3fYbb7yRkpISSktLmTlzJgCVlZWMGzeO0tJSRowYwSuvvMLSpUv57Gc/W/e66dOnU1FRAdQ+afO6665j9OjR3HPPPfzHf/wHI0eOpLS0lLPPPputW7cCsGHDBiZPnkxpaSmlpaU89thjfPvb3+amm26qO+5VV13FrFmzMunzQftoUkmSmrJgwQImTJjA8ccfT8+ePXn66afZsGEDCxYs4IknnuDQQw/l7bffBmDq1KnMnDmTyZMns23bNnbt2sX69ev3ePyuXbvyyCOPALBp0ya+/OUvA3D11Vdz++23M2PGDK644grGjBnDfffdR01NDVu2bOGYY47hrLPO4u/+7u/YtWsX8+bN48knn8ykzwa2JGm//b9fvcDq6j9meszBx3TnO2d+fI9l5s6dy9e+9jUALrjgAubOncuuXbu45JJLOPTQQwHo2bMnmzdv5vXXX2fy5MlAbRA3x/nnn1+3vGrVKq6++mreeecdtmzZwqc//WkAHnroIe666y4AioqK6NGjBz169KBXr14888wzbNiwgeHDh9OrV699ewOaYGBLknJl06ZNPPTQQ6xatYqIoKamhojg7LPP3u3u65RSo8fo1KkTu3btqltv+BGrww47rG75i1/8IgsWLKC0tJSKigqWLl26x/Z96UtfoqKigjfffJO/+Zu/2cfeNc3AliTtt72NhA+Ee++9l4suuohbbrmlbtuYMWPo2bMnd9xxB5///OfrTon37NmT4uJiFixYwOc+9zm2b99OTU0N/fv3Z/Xq1Wzfvp1t27axePFiRo9u9JlebN68maOPPpqdO3dy9913c+yxxwIwduxYfvKTn/C1r32Nmpoa/vSnP9G9e3cmT57MNddcw86dO/nZz36WWb+96UySlCtz586tO8X9gbPPPpvq6momTpxIWVkZw4YN4wc/+AEAc+bMYdasWQwdOpRTTjmFN998k759+3LeeecxdOhQpk6dyvDhw5us7/rrr+ekk05i/PjxnHDCCXXbb7rpJpYsWUJJSQmf+MQneOGFFwDo0qULp59+Ouedd16md5hHU6cL2lpZWVlyPmxJOvisWbOGE088sa2bcdDatWsXI0aM4J577mHQoEFNlmvsfYyIp1JKZY2Vd4QtSVJGVq9ezXHHHcfYsWP3GNb7w2vYkiRlZPDgwbz66qsH5NiOsCVJygEDW5KkHDCwJUnKAQNbkqQcMLAlSblTVFTEsGHDGDJkCGeeeSbvvPNOpsevqKhg+vTpAFx77bV1n+luSwa2JCl3DjnkEFauXMmqVavo2bMns2fPbusmHXAGtiQp104++WRef/31uvXvf//7jBw5kqFDh/Kd73ynbvtdd93F0KFDKS0t5cILLwTgV7/6FSeddBLDhw9n3LhxbNiwodXb31x+DluSlFs1NTUsXryYadOmAfDggw/y8ssv8+STT5JSYuLEiTz88MP06tWL7373uzz66KP07t27burN0aNHs2zZMiKC2267jRtvvJEf/vCHbdmlJhnYkqT998BMePP5bI95VAmc8a97LPLee+8xbNgw1q1bxyc+8QnGjx8P1Ab2gw8+WPds8C1btvDyyy/z7LPPcs4559C7d2+gdupNgKqqKs4//3zeeOMNduzYwcCBA7PtS4Y8JS5Jyp0PrmG/9tpr7Nixo+4adkqJb33rW6xcuZKVK1dSWVnJtGnTSCntNvUmwIwZM5g+fTrPP/88t9xyy27TbB5MHGFLkvbfXkbCB1qPHj2YNWsWkyZN4qtf/Sqf/vSn+fa3v83UqVPp1q0br7/+Op07d2bs2LFMnjyZr3/96/Tq1atu6s133323brrMO++8s037sjcGtiQp14YPH05paSnz5s3jwgsvZM2aNZx88skAdOvWjZ/+9Kd8/OMf56qrrmLMmDEUFRUxfPhwKioquPbaazn33HM59thjGTVqFL/97W/buDdNc3pNSdI+cXrNbDi9piRJ7ZCBLUlSDhjYkiTlgIEtSVIOGNiSJOWAgS1JUg5kEtgRcUdEvBURq5rYHxExKyIqI+K5iBiRRb2SpI6p/vSa5557Llu3bm3xMVesWMEVV1zR5P7q6mrOOeecFtezv7IaYVcAE/aw/wxgUOHrUuAnGdUrSeqA6k+v2aVLF26++eYP7U8psWvXrn06ZllZGbNmzWpy/zHHHMO99967X+3NQiaBnVJ6GHh7D0UmAXelWsuAj0bE0VnULUnq2P7yL/+SyspK1q1bx4knnsjll1/OiBEjWL9+PQ8++CAnn3wyI0aM4Nxzz2XLli0ALF++nFNOOYXS0lLKy8vZvHkzS5cu5bOf/SwA//u//8uwYcMYNmwYw4cPZ/Pmzaxbt44hQ4YAsG3bNi655BJKSkoYPnw4S5YsAaCiooKzzjqLCRMmMGjQIP7hH/4hs3621jXsY4H19darCtskSdpv77//Pg888AAlJSUArF27losuuohnnnmGww47jBtuuIFFixbx9NNPU1ZWxo9+9CN27NjB+eefz0033cSzzz7LokWLOOSQQz503B/84AfMnj2blStX8pvf/Ga3/R9MNvL8888zd+5cLr744rqJQ1auXMn8+fN5/vnnmT9/PuvXrycLrfUs8d2nSIHdnokaEZdSe8qcfv36Heg2SZJa6HtPfo8X334x02Oe0PME/rH8H/dY5oPpNaF2hD1t2jSqq6vp378/o0aNAmDZsmWsXr2aU089FYAdO3Zw8skns3btWo4++mhGjhwJQPfu3Xc7/qmnnsqVV17J1KlTOeussyguLv7Q/kceeYQZM2bUtveEE+jfvz8vvfQSAGPHjqVHjx4ADB48mNdee42+ffvu79tRp7UCuwqo39pioLphoZTSrcCtUPss8dZpmiQpbz64ht3QYYcdVrecUmL8+PHMnTv3Q2Wee+65RqfarG/mzJn89V//Nffffz+jRo1i0aJFdO3a9UPHbspHPvKRuuWioiLef//9vfanOVorsBcC0yNiHnAS8G5K6Y1WqluSdIDsbSTclkaNGsXf/u3fUllZyXHHHcfWrVupqqrihBNOoLq6muXLlzNy5Eg2b9682ynvV155hZKSEkpKSnj88cd58cUX60b0AKeddhp33303f/VXf8VLL73E7373Oz72sY/x9NNPH7D+ZBLYETEX+CTQOyKqgO8AnQFSSjcD9wOfASqBrcAlWdQrSVJT+vTpQ0VFBVOmTGH79u0A3HDDDRx//PHMnz+fGTNm8N5773HIIYewaNGiD732xz/+MUuWLKGoqIjBgwdzxhln8MYb/zfOvPzyy7nssssoKSmhU6dOVFRUfGhkfSA4vaYkaZ84vWY2nF5TkqR2yMCWJCkHDGxJknLAwJYkKQcMbEmScsDAliQpBwxsSZJywMCWJOXOB/Nhf/C1bt06Nm3axOmnn063bt2YPn36Xo9xxx13UFJSwtChQxkyZAi//OUvW6Hl+6+1Hk0qSVJmGnuW+J/+9Ceuv/56Vq1axapVq/b4+qqqKr773e/y9NNP06NHD7Zs2cLGjRtb1KaamhqKiopadIw9cYQtSWoXDjvsMEaPHv2hSTqa8tZbb3H44YfTrVs3ALp168bAgQMBqKysZNy4cZSWljJixAheeeUVUkp885vfZMiQIZSUlDB//nwAli5dyumnn87nP//5uik+f/rTn1JeXs6wYcP4yle+Qk1NTSb9c4QtSdpvb/7zP7N9TbbTa37kxBM46p/+aY9l6k+vOXDgQO677759qqO0tJQjjzySgQMHMnbsWM466yzOPPNMAKZOncrMmTOZPHky27ZtY9euXfziF79g5cqVPPvss/z+979n5MiRnHbaaQA8+eSTrFq1ioEDB7JmzRrmz5/Po48+SufOnbn88su5++67ueiii/bjnfgwA1uSlDtNTa/ZXEVFRfz3f/83y5cvZ/HixXz961/nqaee4hvf+Aavv/46kydPBqgbrT/yyCNMmTKFoqIijjzySMaMGcPy5cvp3r075eXldaPzxYsX89RTT9XNtf3ee+9xxBFHtLC3tQxsSdJ+29tI+GAWEZSXl1NeXs748eO55JJLuPLKKxstu6eJshrOwX3xxRfzL//yL5m312vYkqQOp7q6+kNzV69cuZL+/fvTvXt3iouLWbBgAQDbt29n69atnHbaacyfP5+amho2btzIww8/THl5+W7HHTt2LPfeey9vvfUWAG+//TavvfZaJm12hC1JajcGDBjAH//4R3bs2MGCBQt48MEHGTx48G7ldu7cyd///d9TXV1N165d6dOnDzfffDMAc+bM4Stf+QrXXHMNnTt35p577mHy5Mk8/vjjlJaWEhHceOONHHXUUbz44oev3w8ePJgbbriBT33qU+zatYvOnTsze/Zs+vfv3+K+OR+2JGmfOB92NpwPW5KkdshT4pKkdu2kk05i+/btH9o2Z86cus9N54WBLUlq15544om2bkImPCUuSdpnB+v9T3mxP++fgS1J2iddu3Zl06ZNhvZ+SimxadOmZj1CtT5PiUuS9klxcTFVVVUtniyjI+vatSvFxcX79BoDW5K0Tzp37lz3KE61Hk+JS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAOZBHZETIiItRFRGREzG9nfLyKWRMQzEfFcRHwmi3olSeooWhzYEVEEzAbOAAYDUyKi4VxmVwM/TykNBy4A/r2l9UqS1JFkMcIuBypTSq+mlHYA84BJDcokoHthuQdQnUG9kiR1GFk8OOVYYH299SrgpAZlrgUejIgZwGHAuAzqlSSpw8hihB2NbGv4gNkpQEVKqRj4DDAnInarOyIujYgVEbHCR95JkvR/sgjsKqBvvfVidj/lPQ34OUBK6XGgK9C74YFSSremlMpSSmV9+vTJoGmSJLUPWQT2cmBQRAyMiC7U3lS2sEGZ3wFjASLiRGoD2yG0JEnN1OLATim9D0wHfg2sofZu8Bci4rqImFgo9g3gyxHxLDAX+GJyXjZJkpotk9m6Ukr3A/c32HZNveXVwKlZ1CVJUkfkk84kScoBA1uSpBwwsCVJygEDW5KkHDCwJUnKAQNbkqQcMLAlScoBA1uSpBwwsCVJygEDW5KkHDCwJUnKAQNbkqQcMLAlScoBA1uSpBwwsCVJygEDW5KkHDCwJUnKAQNbkqQcMLAlScoBA1uSpBwwsCVJygEDW5KkHDCwJUnKAQNbkqQcMLAlScoBA1uSpBwwsCVJygEDW5KkHDCwJUnKgUwCOyImRMTaiKiMiJlNlDkvIlZHxAsR8bMs6pUkqaPo1NIDREQRMBsYD1QByyNiYUppdb0yg4BvAaemlP4QEUe0tF5JkjqSLEbY5UBlSunVlNIOYB4wqUGZLwOzU0p/AEgpvZVBvZIkdRhZBPaxwPp661WFbfUdDxwfEY9GxLKImNDYgSLi0ohYERErNm7cmEHTJElqH7II7GhkW2qw3gkYBHwSmALcFhEf3e1FKd2aUipLKZX16dMng6ZJktQ+ZBHYVUDfeuvFQHUjZX6ZUtqZUvotsJbaAJckSc2QRWAvBwZFxMCI6AJcACxsUGYBcDpARPSm9hT5qxnULUlSh9DiwE4pvQ9MB34NrAF+nlJ6ISKui4iJhWK/BjZFxGpgCfDNlNKmltYtSVJHESk1vNx8cCgrK0srVqxo62ZIktRqIuKplFJZY/t80pkkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUA5kEdkRMiIi1EVEZETP3UO6ciEgRUZZFvZIkdRQtDuyIKAJmA2cAg4EpETG4kXKHA1cAT7S0TkmSOposRtjlQGVK6dWU0g5gHjCpkXLXAzcC2zKoU5KkDiWLwD4WWF9vvaqwrU5EDAf6ppT+c08HiohLI2JFRKzYuHFjBk2TJKl9yCKwo5FtqW5nxJ8B/wZ8Y28HSindmlIqSymV9enTJ4OmSZLUPmQR2FVA33rrxUB1vfXDgSHA0ohYB4wCFnrjmSRJzZdFYC8HBkXEwIjoAlwALPxgZ0rp3ZRS75TSgJTSAGAZMDGltCKDuiVJ6hBaHNgppfeB6cCvgTXAz1NKL0TEdRExsaXHlyRJ0CmLg6SU7gfub7DtmibKfjKLOiVJ6kh80pkkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUA5kEdkRMiIi1EVEZETMb2X9lRKyOiOciYnFE9M+iXkmSOooWB3ZEFAGzgTOAwcCUiBjcoNgzQFlKaShwL3BjS+uVJKkjyWKEXQ5UppReTSntAOYBk+oXSCktSSltLawuA4ozqFeSpA4ji8A+Flhfb72qsK0p04AHMqhXkqQOo1MGx4hGtqVGC0Z8ASgDxjSx/1LgUoB+/fpl0DRJktqHLEbYVUDfeuvFQHXDQhExDrgKmJhS2t7YgVJKt6aUylJKZX369MmgaZIktQ9ZBPZyYFBEDIyILsAFwML6BSJiOHALtWH9VgZ1SpLUobQ4sFNK7wPTgV8Da4Cfp5ReiIjrImJiodj3gW7APRGxMiIWNnE4SZLUiCyuYZNSuh+4v8G2a+otj8uiHkmSOiqfdCZJUg4Y2JIk5YCBLUlSDhjYkiTlgIEtSVIOGNiSJOWAgS1JUg4Y2JIk5YCBLUlSDhjYkiTlgIEtSVIOGNiSJOWAgS1JUg4Y2JIk5YCBLUlSDhjYkiTlgIEtSVIOGNiSJOWAgS1JUg4Y2JIk5YCBLUlSDhjYkiTlgIEtSVIOGNiSJOWAgS1JUg4Y2JIk5YCBLUlSDhjYkiTlQCaBHRETImJtRFRGxMxG9n8kIuYX9j8REQOyqFeSpI6ixYEdEUXAbOAMYDAwJSIGNyg2DfhDSuk44N+A77W0XkmSOpIsRtjlQGVK6dWU0g5gHjCpQZlJwJ2F5XuBsRERGdQtSVKHkEVgHwusr7deVdjWaJmU0vvAu0CvDOqWJKlDyCKwGxspp/0oQ0RcGhErImLFxo0bM2iaJEntQxaBXQX0rbdeDFQ3VSYiOgE9gLcbHiildGtKqSylVNanT58MmiZJUvuQRWAvBwZFxMCI6AJcACxsUGYhcHFh+RzgoZTSbiNsSZLUuE4tPUBK6f2ImA78GigC7kgpvRAR1wErUkoLgduBORFRSe3I+oKW1itJUkfS4sAGSCndD9zfYNs19Za3AedmUZckSR2RTzqTJCkHDGxJknLAwJYkKQcMbEmScsDAliQpBwxsSZJywMCWJCkHDGxJknLAwJYkKQcMbEmScsDAliQpBwxsSZJywMCWJCkHDGxJknLAwJYkKQcMbEmScsDAliQpBwxsSZJywMCWJCkHDGxJknLAwJYkKQcMbEmScsDAliQpBwxsSZJywMCWJCkHDGxJknLAwJYkKQcMbEmScqBFgR0RPSPifyLi5cL3P2+kzLCIeDwiXoiI5yLi/JbUKUlSR9TSEfZMYHFKaRCwuLDe0FbgopTSx4EJwI8j4qMtrFeSpA6lpYE9CbizsHwn8LmGBVJKL6WUXi4sVwNvAX1aWK8kSR1KSwP7yJTSGwCF70fsqXBElANdgFdaWK8kSR1Kp70ViIhFwFGN7LpqXyqKiKOBOcDFKaVdTZS5FLgUoF+/fvtyeEmS2rW9BnZKaVxT+yJiQ0QcnVJ6oxDIbzVRrjvwX8DVKaVle6jrVuBWgLKysrS3tkmS1FG09JT4QuDiwvLFwC8bFoiILsB9wF0ppXtaWJ8kSR1SSwP7X4HxEfEyML6wTkSURcRthTLnAacBX4yIlYWvYS2sV5KkDiVSOjjPPJeVlaUVK1a0dTMkSWo1EfFUSqmssX0+6UySpBw4aEfYEbEReK0ZRXsDvz/AzTnY2OeOwT53DPa5Y2hun/unlBp9VslBG9jNFRErmjp90F7Z547BPncM9rljyKLPnhKXJCkHDGxJknKgPQT2rW3dgDZgnzsG+9wx2OeOocV9zv01bEmSOoL2MMKWJKndy11gR0TPiPifiHi58P3PGykzLCIej4gXIuK5iDi/LdraUhExISLWRkRlROw213hEfCQi5hf2PxERA1q/ldlqRp+vjIjVhZ/r4ojo3xbtzNLe+lyv3DkRkSIi93fXNqfPEXFe4Wf9QkT8rLXbmLVm/G73i4glEfFM4ff7M23RzqxExB0R8VZErGpif0TErML78VxEjGjtNmatGX2eWujrcxHxWESU7lMFKaVcfQE3AjMLyzOB7zVS5nhgUGH5GOAN4KNt3fZ97GcRtdOQ/gW1U5I+CwxuUOZy4ObC8gXA/LZudyv0+XTg0MLyVztCnwvlDgceBpYBZW3d7lb4OQ8CngH+vLB+RFu3uxX6fCvw1cLyYGBdW7e7hX0+DRgBrGpi/2eAB4AARgFPtHWbW6HPp9T7nT5jX/ucuxE2MAm4s7B8J/C5hgVSSi+llF4uLFdTO4tYox9EP4iVA5UppVdTSjuAedT2vb7678W9wNiIiFZsY9b22ueU0pKU0tbC6jKguJXbmLXm/JwBrqf2P6vbWrNxB0hz+vxlYHZK6Q8AKaVGZwLMkeb0OQHdC8s9gOpWbF/mUkoPA2/vocgkaieFSql2FsePFmZ9zK299Tml9NgHv9Psx9+vPAb2kSmlNwAK34/YU+GIKKf2f7SvtELbsnQssL7eelVhW6NlUkrvA+8CvVqldQdGc/pc3zRq/4eeZ3vtc0QMB/qmlP6zNRt2ADXn53w8cHxEPBoRyyJiQqu17sBoTp+vBb4QEVXA/cCM1mlam9nXf+/tzT7//drrfNhtISIWAUc1suuqfTzO0cAc4OKU0q4s2taKGhspN7ylvzll8qTZ/YmILwBlwJgD2qIDb499jog/A/4N+GJrNagVNOfn3Ina0+KfpHYU8puIGJJSeucAt+1AaU6fpwAVKaUfRsTJwJxCn/P2t6u52tvfr2aLiNOpDezR+/K6gzKwU0rjmtoXERsi4uiU0huFQG70VFlEdAf+C7i6cLolb6qAvvXWi9n9FNkHZaoiohO1p9H2dArqYNecPhMR46j9z9uYlNL2VmrbgbK3Ph8ODAGWFq52HAUsjIiJKaW8TmfX3N/tZSmlncBvI2IttQG+vHWamLnm9HkaMAEgpfR4RHSl9vnTeb8c0JRm/XtvbyJiKHAbcEZKadO+vDaPp8QXAhcXli8GftmwQER0Ae6j9vrIPa3YtiwtBwZFxMBCfy6gtu/11X8vzgEeSoW7GXJqr30unB6+BZjYDq5rwl76nFJ6N6XUO6U0IKU0gNrrXnkOa2je7/YCam8wJCJ6U3uK/NVWbWW2mtPn3wFjASLiRKArsLFVW9m6FgIXFe4WHwW8+8HlzvYqIvoBvwAuTCm9tM8HaOu76vbjLrxewGLg5cL3noXtZcBtheUvADuBlfW+hrV12/ejr58BXqL2+vtVhW3XUfsHG2r/Qd8DVAJPAn/R1m1uhT4vAjbU+7kubOs2H+g+Nyi7lJzfJd7Mn3MAPwJWA88DF7R1m1uhz4OBR6m9g3wl8Km2bnML+zuX2k/o7KR2ND0NuAy4rN7PeHbh/Xi+nfxe763PtwF/qPf3a8W+HN8nnUmSlAN5PCUuSVKHY2BLkpQDBrYkSTlgYEuSlAMGtiRJOWBgS5KUAwa2JEk5YGBLkpQD/x+GlauYgyQhuQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM\n",
       "2                k4  dtree      kNN     dtree      kNN  dtree  SVM  SVM    SVM\n",
       "3                k3    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #k=2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='k2'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [Classifier, TP, FP, TN, FN, Accuracy, Recall, Precision, F1_Score]\n",
       "Index: []"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM\n",
       "2                k4  dtree      kNN     dtree      kNN  dtree  SVM  SVM    SVM\n",
       "3                k3    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN\n",
       "4                k2    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #L1-based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='L1'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kNN</td>\n",
       "      <td>4321</td>\n",
       "      <td>1051</td>\n",
       "      <td>4154</td>\n",
       "      <td>1307</td>\n",
       "      <td>0.782332</td>\n",
       "      <td>0.767768</td>\n",
       "      <td>0.804356</td>\n",
       "      <td>0.785636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>gnb</td>\n",
       "      <td>4133</td>\n",
       "      <td>2120</td>\n",
       "      <td>3085</td>\n",
       "      <td>1495</td>\n",
       "      <td>0.666297</td>\n",
       "      <td>0.734364</td>\n",
       "      <td>0.660963</td>\n",
       "      <td>0.695733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>dtree</td>\n",
       "      <td>3556</td>\n",
       "      <td>1211</td>\n",
       "      <td>3994</td>\n",
       "      <td>2072</td>\n",
       "      <td>0.696945</td>\n",
       "      <td>0.631841</td>\n",
       "      <td>0.745962</td>\n",
       "      <td>0.684175</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>SVM</td>\n",
       "      <td>3787</td>\n",
       "      <td>1545</td>\n",
       "      <td>3660</td>\n",
       "      <td>1841</td>\n",
       "      <td>0.687437</td>\n",
       "      <td>0.672886</td>\n",
       "      <td>0.710240</td>\n",
       "      <td>0.691058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>nn</td>\n",
       "      <td>4227</td>\n",
       "      <td>2618</td>\n",
       "      <td>2587</td>\n",
       "      <td>1401</td>\n",
       "      <td>0.629004</td>\n",
       "      <td>0.751066</td>\n",
       "      <td>0.617531</td>\n",
       "      <td>0.677784</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>lr</td>\n",
       "      <td>4172</td>\n",
       "      <td>2574</td>\n",
       "      <td>2631</td>\n",
       "      <td>1456</td>\n",
       "      <td>0.627989</td>\n",
       "      <td>0.741294</td>\n",
       "      <td>0.618441</td>\n",
       "      <td>0.674317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>sgd</td>\n",
       "      <td>4567</td>\n",
       "      <td>3123</td>\n",
       "      <td>2082</td>\n",
       "      <td>1061</td>\n",
       "      <td>0.613773</td>\n",
       "      <td>0.811478</td>\n",
       "      <td>0.593888</td>\n",
       "      <td>0.685839</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  F1_Score\n",
       "3         kNN  4321  1051  4154  1307  0.782332  0.767768   0.804356  0.785636\n",
       "6         gnb  4133  2120  3085  1495  0.666297  0.734364   0.660963  0.695733\n",
       "11      dtree  3556  1211  3994  2072  0.696945  0.631841   0.745962  0.684175\n",
       "16        SVM  3787  1545  3660  1841  0.687437  0.672886   0.710240  0.691058\n",
       "21         nn  4227  2618  2587  1401  0.629004  0.751066   0.617531  0.677784\n",
       "26         lr  4172  2574  2631  1456  0.627989  0.741294   0.618441  0.674317\n",
       "31        sgd  4567  3123  2082  1061  0.613773  0.811478   0.593888  0.685839"
      ]
     },
     "execution_count": 67,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>L1</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM\n",
       "2                k4  dtree      kNN     dtree      kNN  dtree  SVM  SVM    SVM\n",
       "3                k3    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN\n",
       "4                k2    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN\n",
       "5                L1    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## #Tree-based"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='Tree'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.drop(['Model'], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>kNN</td>\n",
       "      <td>4426</td>\n",
       "      <td>1278</td>\n",
       "      <td>3927</td>\n",
       "      <td>1202</td>\n",
       "      <td>0.771070</td>\n",
       "      <td>0.786425</td>\n",
       "      <td>0.775947</td>\n",
       "      <td>0.781151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>gnb</td>\n",
       "      <td>3897</td>\n",
       "      <td>2001</td>\n",
       "      <td>3204</td>\n",
       "      <td>1731</td>\n",
       "      <td>0.655497</td>\n",
       "      <td>0.692431</td>\n",
       "      <td>0.660732</td>\n",
       "      <td>0.676210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>SVM</td>\n",
       "      <td>4581</td>\n",
       "      <td>2868</td>\n",
       "      <td>2337</td>\n",
       "      <td>1047</td>\n",
       "      <td>0.638604</td>\n",
       "      <td>0.813966</td>\n",
       "      <td>0.614982</td>\n",
       "      <td>0.700619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>nn</td>\n",
       "      <td>3734</td>\n",
       "      <td>2260</td>\n",
       "      <td>2945</td>\n",
       "      <td>1894</td>\n",
       "      <td>0.616542</td>\n",
       "      <td>0.663468</td>\n",
       "      <td>0.622956</td>\n",
       "      <td>0.642574</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>lr</td>\n",
       "      <td>3870</td>\n",
       "      <td>2458</td>\n",
       "      <td>2747</td>\n",
       "      <td>1758</td>\n",
       "      <td>0.610819</td>\n",
       "      <td>0.687633</td>\n",
       "      <td>0.611568</td>\n",
       "      <td>0.647374</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>sgd</td>\n",
       "      <td>4048</td>\n",
       "      <td>2980</td>\n",
       "      <td>2225</td>\n",
       "      <td>1580</td>\n",
       "      <td>0.579064</td>\n",
       "      <td>0.719261</td>\n",
       "      <td>0.575982</td>\n",
       "      <td>0.639697</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  F1_Score\n",
       "4         kNN  4426  1278  3927  1202  0.771070  0.786425   0.775947  0.781151\n",
       "7         gnb  3897  2001  3204  1731  0.655497  0.692431   0.660732  0.676210\n",
       "12      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868  0.678095\n",
       "17        SVM  4581  2868  2337  1047  0.638604  0.813966   0.614982  0.700619\n",
       "22         nn  3734  2260  2945  1894  0.616542  0.663468   0.622956  0.642574\n",
       "27         lr  3870  2458  2747  1758  0.610819  0.687633   0.611568  0.647374\n",
       "32        sgd  4048  2980  2225  1580  0.579064  0.719261   0.575982  0.639697"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[5:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Volumes/BECCA/Project/Data/Result/%s/By_model/%s.png' %(Country, ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [],
   "source": [
    "opt={'Feature Selection':ft_name}\n",
    "for i in list(result.index):\n",
    "    for j in result.columns:\n",
    "        if j != 'Classifier':\n",
    "            if result[j].max() == result[j][i]:\n",
    "                opt[j]=result.Classifier[i]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>TN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "      <th>FN</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>Recall</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>nn</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>L1</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "      <td>sgd</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Tree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>dtree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>sgd</td>\n",
       "      <td>SVM</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection     TN Accuracy Precision F1_Score     FN   TP   FP Recall\n",
       "0                k6    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "1                k5  dtree      kNN     dtree      kNN  dtree  SVM   nn    SVM\n",
       "2                k4  dtree      kNN     dtree      kNN  dtree  SVM  SVM    SVM\n",
       "3                k3    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN\n",
       "4                k2    NaN      NaN       NaN      NaN    NaN  NaN  NaN    NaN\n",
       "5                L1    kNN      kNN       kNN      kNN  dtree  sgd  sgd    sgd\n",
       "6              Tree  dtree      kNN     dtree      kNN  dtree  SVM  sgd    SVM"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result=Ana_result.append(opt, ignore_index=True)\n",
    "Ana_result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Feature Selection</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>k6</td>\n",
       "      <td>kNN</td>\n",
       "      <td>sgd</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>k5</td>\n",
       "      <td>kNN</td>\n",
       "      <td>SVM</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>k4</td>\n",
       "      <td>kNN</td>\n",
       "      <td>SVM</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>k3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>k2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>L1</td>\n",
       "      <td>kNN</td>\n",
       "      <td>sgd</td>\n",
       "      <td>kNN</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>Tree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>SVM</td>\n",
       "      <td>dtree</td>\n",
       "      <td>kNN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Feature Selection Accuracy Recall Precision F1_Score\n",
       "0                k6      kNN    sgd       kNN      kNN\n",
       "1                k5      kNN    SVM     dtree      kNN\n",
       "2                k4      kNN    SVM     dtree      kNN\n",
       "3                k3      NaN    NaN       NaN      NaN\n",
       "4                k2      NaN    NaN       NaN      NaN\n",
       "5                L1      kNN    sgd       kNN      kNN\n",
       "6              Tree      kNN    SVM     dtree      kNN"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Ana_result[['Feature Selection', 'Accuracy', 'Recall', 'Precision', 'F1_Score']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "ft_name='Tree'\n",
    "result=data.loc[data['Model'] == ft_name]\n",
    "result=result.reset_index()\n",
    "for i in range(4,8):\n",
    "    result=result.drop(i)\n",
    "result=result.drop(['index'],axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Model</th>\n",
       "      <th>Classifier</th>\n",
       "      <th>TP</th>\n",
       "      <th>FP</th>\n",
       "      <th>TN</th>\n",
       "      <th>FN</th>\n",
       "      <th>Accuracy</th>\n",
       "      <th>Recall</th>\n",
       "      <th>Precision</th>\n",
       "      <th>F1_Score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Tree</td>\n",
       "      <td>kNN</td>\n",
       "      <td>4426</td>\n",
       "      <td>1278</td>\n",
       "      <td>3927</td>\n",
       "      <td>1202</td>\n",
       "      <td>0.771070</td>\n",
       "      <td>0.786425</td>\n",
       "      <td>0.775947</td>\n",
       "      <td>0.781151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Tree</td>\n",
       "      <td>gnb</td>\n",
       "      <td>3897</td>\n",
       "      <td>2001</td>\n",
       "      <td>3204</td>\n",
       "      <td>1731</td>\n",
       "      <td>0.655497</td>\n",
       "      <td>0.692431</td>\n",
       "      <td>0.660732</td>\n",
       "      <td>0.676210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Tree</td>\n",
       "      <td>dtree</td>\n",
       "      <td>3303</td>\n",
       "      <td>811</td>\n",
       "      <td>4394</td>\n",
       "      <td>2325</td>\n",
       "      <td>0.710514</td>\n",
       "      <td>0.586887</td>\n",
       "      <td>0.802868</td>\n",
       "      <td>0.678095</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Tree</td>\n",
       "      <td>SVM</td>\n",
       "      <td>4581</td>\n",
       "      <td>2868</td>\n",
       "      <td>2337</td>\n",
       "      <td>1047</td>\n",
       "      <td>0.638604</td>\n",
       "      <td>0.813966</td>\n",
       "      <td>0.614982</td>\n",
       "      <td>0.700619</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Model Classifier    TP    FP    TN    FN  Accuracy    Recall  Precision  \\\n",
       "0  Tree        kNN  4426  1278  3927  1202  0.771070  0.786425   0.775947   \n",
       "1  Tree        gnb  3897  2001  3204  1731  0.655497  0.692431   0.660732   \n",
       "2  Tree      dtree  3303   811  4394  2325  0.710514  0.586887   0.802868   \n",
       "3  Tree        SVM  4581  2868  2337  1047  0.638604  0.813966   0.614982   \n",
       "\n",
       "   F1_Score  \n",
       "0  0.781151  \n",
       "1  0.676210  \n",
       "2  0.678095  \n",
       "3  0.700619  "
      ]
     },
     "execution_count": 103,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "\n",
    "x=result.Classifier\n",
    "for col in result.columns[6:10]:\n",
    "    y=result[col]\n",
    "    plt.plot(x,y, label=col)\n",
    "\n",
    "plt.legend()\n",
    "plt.savefig('/Users/becca/Desktop/%s.png' %(ft_name))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}