demo_featureSelector.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "outputs": [],
   "source": [
    "import warnings\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "\n",
    "warnings.filterwarnings('ignore')\n",
    "%matplotlib inline\n",
    "sns.set(rc={'figure.figsize':(14,9)})"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Lade in Spark-Scala aufbereitete Tabelle\n",
    "https://a100.sbb.ch:8974/notebooks/users/u230261/STA1911/statAnalysis.snb.ipynb?read_only=1"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "311b0281",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv (r'C:\\devsbb\\lab_dlfw.stats_sta1911_v2.csv',\n",
    "                  delimiter=\";\",\n",
    "                  dtype={'profil_typ': object}, # Sonsts wirds als Exponential-Zahl interpretiert (60E1 = 600.0)\n",
    "                  usecols=[\"channel\",\"start_mm\",\"slope_mm_yr\",\"TgGewicht\",\"curve\",\"lastklasse\",\"profil_typ\",\"schw_art\",\"radiusklasse\",\"stat_test_p_min\",\"n_measurements\",\"time_span_yrs\"])\n",
    "    #.dropna(inplace=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "outputs": [
    {
     "data": {
      "text/plain": "2049687"
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df.index)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "outputs": [],
   "source": [
    "## initiales Filtern : nur Prognose-Äste mit >= 4 samples über min 6 Monate Zeitspanne und bestandenen stat. Tests\n",
    "df = df[(df.n_measurements >= 4) & (df.time_span_yrs > 0.5) & (df.stat_test_p_min > 0.05/3.0) & (df.channel == \"lonled1_ex_std200m\")] # Division durch 3 : Bonferroni-Korrektur"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "outputs": [
    {
     "data": {
      "text/plain": "532331"
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df.index) #Die Hälfte aller Punkte fliegt hier also weg...."
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "### Playground Feature-Analysis"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "37a61bcc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 864x648 with 2 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "corrmat = df.corr()\n",
    "f, ax = plt.subplots(figsize=(12, 9))\n",
    "sns.heatmap(corrmat, vmax=.8, square=True);"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "Feature-Importance"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "outputs": [],
   "source": [
    "from sklearn import preprocessing\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.linear_model import LinearRegression"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "outputs": [],
   "source": [
    "x = df[['curve','TgGewicht','schw_art','profil_typ']] ## Auswahl der zu prüfenden Features\n",
    "y = df['slope_mm_yr']"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "outputs": [],
   "source": [
    "## One-Hot encoder für Schwellen-Art\n",
    "discarded = pd.get_dummies(x['schw_art'],prefix = 'schw_art')\n",
    "x = pd.concat([x.drop(\"schw_art\",1), discarded], axis = 1)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "outputs": [],
   "source": [
    "## One-Hot encoder für Profil-Typ\n",
    "discarded = pd.get_dummies(x['profil_typ'],prefix = 'profil_typ')\n",
    "x = pd.concat([x.drop(\"profil_typ\",1), discarded], axis = 1)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "6d792800",
   "metadata": {
    "pycharm": {
     "name": "#%%\n"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feature ranking:\n",
      "1. feature 1 (0.606639)\n",
      "2. feature 0 (0.348424)\n",
      "3. feature 7 (0.009739)\n",
      "4. feature 3 (0.009015)\n",
      "5. feature 4 (0.008124)\n",
      "6. feature 6 (0.006666)\n",
      "7. feature 2 (0.006662)\n",
      "8. feature 5 (0.004731)\n"
     ]
    },
    {
     "data": {
      "text/plain": "<Figure size 1008x504 with 1 Axes>",
      "image/png": "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\n"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "rfr = RandomForestRegressor()\n",
    "f_rfr = rfr.fit(x, y)\n",
    "importances = f_rfr.feature_importances_\n",
    "std = np.std([tree.feature_importances_ for tree in rfr.estimators_],\n",
    "             axis=0)\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "# Print the feature ranking\n",
    "print(\"Feature ranking:\")\n",
    "\n",
    "for f in range(x.shape[1]):\n",
    "    print(\"%d. feature %d (%f)\" % (f + 1, indices[f], importances[indices[f]]))\n",
    "\n",
    "# Plot the feature importances of the forest\n",
    "\n",
    "plt.figure(1, figsize=(14, 7))\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(x.shape[1]), importances[indices],\n",
    "       color=\"g\", yerr=std[indices], align=\"center\")\n",
    "plt.xticks(range(x.shape[1]), x.columns[indices],rotation=90)\n",
    "plt.xlim([-1, x.shape[1]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "source": [
    "Library von WillKoehrsen"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% md\n"
    }
   }
  },
  {
   "cell_type": "raw",
   "source": [
    "https://github.com/WillKoehrsen/feature-selector/blob/master/Feature%20Selector%20Usage.ipynb"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%% raw\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "outputs": [],
   "source": [
    "from feature_selector import FeatureSelector\n",
    "\n",
    "import pandas as pd"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "outputs": [
    {
     "data": {
      "text/plain": "    slope_mm_yr    curve  TgGewicht schw_art profil_typ\n7      0.019903 -0.00099    43919.0     holz       54E2\n9      0.038722 -0.00099    43919.0     holz       54E2\n11     0.006086 -0.00099    47414.0     holz       54E2\n17     0.038573 -0.00099    48232.0     holz       54E2\n19     0.011878 -0.00099    41039.0     holz       54E2",
      "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>slope_mm_yr</th>\n      <th>curve</th>\n      <th>TgGewicht</th>\n      <th>schw_art</th>\n      <th>profil_typ</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>7</th>\n      <td>0.019903</td>\n      <td>-0.00099</td>\n      <td>43919.0</td>\n      <td>holz</td>\n      <td>54E2</td>\n    </tr>\n    <tr>\n      <th>9</th>\n      <td>0.038722</td>\n      <td>-0.00099</td>\n      <td>43919.0</td>\n      <td>holz</td>\n      <td>54E2</td>\n    </tr>\n    <tr>\n      <th>11</th>\n      <td>0.006086</td>\n      <td>-0.00099</td>\n      <td>47414.0</td>\n      <td>holz</td>\n      <td>54E2</td>\n    </tr>\n    <tr>\n      <th>17</th>\n      <td>0.038573</td>\n      <td>-0.00099</td>\n      <td>48232.0</td>\n      <td>holz</td>\n      <td>54E2</td>\n    </tr>\n    <tr>\n      <th>19</th>\n      <td>0.011878</td>\n      <td>-0.00099</td>\n      <td>41039.0</td>\n      <td>holz</td>\n      <td>54E2</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train = df[['slope_mm_yr','curve','TgGewicht','schw_art','profil_typ']]\n",
    "train_labels = train['slope_mm_yr']\n",
    "train.head()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "outputs": [],
   "source": [
    "train = train.drop(columns = ['slope_mm_yr'])"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "outputs": [],
   "source": [
    "fs = FeatureSelector(data = train, labels = train_labels)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 features with a single unique value.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "fs.identify_single_unique()"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "outputs": [
    {
     "data": {
      "text/plain": "[]"
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "single_unique = fs.ops['single_unique']\n",
    "single_unique"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "outputs": [
    {
     "data": {
      "text/plain": "            nunique\nschw_art          3\nprofil_typ        3\nTgGewicht      8714\ncurve         69317",
      "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>nunique</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>schw_art</th>\n      <td>3</td>\n    </tr>\n    <tr>\n      <th>profil_typ</th>\n      <td>3</td>\n    </tr>\n    <tr>\n      <th>TgGewicht</th>\n      <td>8714</td>\n    </tr>\n    <tr>\n      <th>curve</th>\n      <td>69317</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fs.unique_stats"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 features with a correlation magnitude greater than 0.98.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "fs.identify_collinear(correlation_threshold=0.98)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Gradient Boosting Model\n",
      "\n"
     ]
    }
   ],
   "source": [
    "fs.identify_zero_importance(task = 'regression', eval_metric = 'l2',\n",
    "                            n_iterations = 10, early_stopping = False)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "one_hot_features = fs.one_hot_features\n",
    "base_features = fs.base_features\n",
    "print('There are %d original features' % len(base_features))\n",
    "print('There are %d one-hot features' % len(one_hot_features))"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "fs.data_all.head(10)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "zero_importance_features = fs.ops['zero_importance']\n",
    "zero_importance_features[10:15]"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "fs.plot_feature_importances(plot_n=8)"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n",
     "is_executing": true
    }
   }
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}