get-results-downstream.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "#!/usr/bin/env python3\n",
    "#DEPENDS: /home/anne/tmpanne/AEM_output/baseline_classreport.json, /home/anne/tmpanne/AEM_output/baseline_classreport.json,\n",
    "#CREATES: None\n",
    "#TITLE: Data cleaner (cleans output of ML algorithms)\n",
    "#DESCRIPTION:Creates a clean data frame with results of ML on different vectorizers, classifiers, and parameter settings.\n",
    "#PIPE: TRUE\n",
    "\n",
    "import sys\n",
    "print(sys.stdin.read().upper())\n",
    "\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import json\n",
    "import logging\n",
    "import pandas as pd\n",
    "\n",
    "def get_data(file):\n",
    "    path_to_data='/home/anne/tmpanne/AEM_output/'\n",
    "\n",
    "    fname_sml = '{}{}'.format(path_to_data, file)\n",
    "    with open(fname_sml) as handle:\n",
    "        class_report =  json.loads(handle.read())\n",
    "\n",
    "    appended_data = []\n",
    "\n",
    "    for i in class_report:\n",
    "        data = pd.DataFrame.from_dict(i)\n",
    "        appended_data.append(data)\n",
    "    df = pd.concat(appended_data)\n",
    "    df['indicators'] = df.index\n",
    "    return df\n",
    "\n",
    "def merge_baseline_embedding(dataset):\n",
    "    df_baseline = get_data('baseline_classreport_dataset_{}_embed_size_large.json'.format(dataset))\n",
    "    df_embeddings = get_data('embeddings_classreport_dataset_{}_embed_size_large.json'.format(dataset))\n",
    "    df = pd.concat([df_baseline , df_embeddings])\n",
    "    return df\n",
    "\n",
    "def clean_data_metrics(dataset):\n",
    "    df = merge_baseline_embedding(dataset)\n",
    "    i = ['precision', 'recall', 'support', 'f1-score']\n",
    "    df[df['indicators'].isin(i)]\n",
    "    df.reset_index(inplace=True)\n",
    "    d = df[df['indicators'].isin(i)].groupby(['classifier', 'model', 'indicators', 'vectorizer']).max()\n",
    "    return d\n",
    "\n",
    "def clean_data_parameters(dataset):\n",
    "    i = ['clf__C', 'clf__fit_intercept', 'clf__loss', 'clf__max_iter','clf__alpha', 'clf__penalty', 'clf__gamma', 'clf__kernel', 'clf__max_features']\n",
    "\n",
    "    df = merge_baseline_embedding(dataset)\n",
    "    df = df[df['indicators'].isin(i)].groupby(['classifier', 'model', 'indicators', 'vectorizer']).max()\n",
    "    df.reset_index(inplace=True)\n",
    "    e = df.groupby(['classifier', 'model','vectorizer']).apply(lambda g: pd.Series(g.parameters.values, index= + g.indicators.astype(str)))\n",
    "\n",
    "    parameters = pd.DataFrame(e)\n",
    "    parameters.reset_index(inplace=True)\n",
    "\n",
    "    parameters['unique_'] = parameters['classifier'] + \"_\" + parameters['model']  + \"_\" +  parameters['vectorizer']\n",
    "\n",
    "    p = parameters.pivot(index='unique_', columns='indicators', values=0)\n",
    "\n",
    "    parameters = parameters.groupby(['classifier', 'model', 'vectorizer']).max()\n",
    "    p.reset_index(inplace=True)\n",
    "    len(p) == len(parameters)\n",
    "    parameters.reset_index(inplace=True)\n",
    "    parameters = pd.merge(p, parameters, on='unique_')\n",
    "    return parameters\n",
    "\n",
    "def get_cleaned_data(dataset):\n",
    "    df = pd.merge(clean_data_metrics(dataset), clean_data_parameters(dataset), on=['classifier','model','vectorizer'], how='left')\n",
    "    df.rename(columns={'index': 'metrics'}, inplace=True)\n",
    "    print(\"..........loaded the data frame........\")\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys  \n",
    "sys.path.append('src/data-processing/')  \n",
    "from datacleaning import * \n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "output_path = 'output/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "..........loaded the data frame........\n"
     ]
    }
   ],
   "source": [
    "b =  get_cleaned_data(\"burscher\")\n",
    "b.head()\n",
    "b.set_index('model', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# select metric of interest\n",
    "df = b[b['metrics'] == 'f1-score']\n",
    "averaging = 'macro avg'\n",
    "#averaging = 'weighted avg'\n",
    "#averaging = 'accuracy'\n",
    "\n",
    "results = df[df.groupby(['model'])[averaging].transform(max) == df[averaging]].sort_values(averaging)\n",
    "results[['classifier', 'vectorizer', averaging , 'clf__alpha', 'clf__fit_intercept', 'clf__gamma', 'clf__kernel',\n",
    "       'clf__loss', 'clf__max_features', 'clf__max_iter', 'clf__penalty']]\n",
    "\n",
    "results.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.catplot(x=\"macro avg\", y=\"model\", edgecolor=\".4\", data=results, kind=\"bar\",  height=8)\n",
    "ax.set_titles(\"{col_name} {col_var}\").despine(left=True)\n",
    "ax.set(xlabel=None, ylabel=None)\n",
    "ax = sns.set_style(\"white\")\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)\n",
    "plt.savefig(\"{}downstream_Burscher.png\".format(output_path),  bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "..........loaded the data frame........\n"
     ]
    }
   ],
   "source": [
    "v =  get_cleaned_data(\"vermeer\")\n",
    "v.head()\n",
    "v.set_index('model', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# select metric of interest\n",
    "df = v[v['metrics'] == 'f1-score']\n",
    "averaging = 'macro avg'\n",
    "\n",
    "results = df[df.groupby(['model'])[averaging].transform(max) == df[averaging]].sort_values(averaging)\n",
    "results[['classifier', 'vectorizer', averaging , 'clf__alpha', 'clf__fit_intercept', 'clf__gamma', 'clf__kernel',\n",
    "       'clf__loss', 'clf__max_features', 'clf__max_iter', 'clf__penalty']]\n",
    "\n",
    "results.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.catplot(x=\"macro avg\", y=\"model\", edgecolor=\".4\", data=results, kind=\"bar\",  height=8)\n",
    "ax.set_titles(\"{col_name} {col_var}\").despine(left=True)\n",
    "ax.set(xlabel=None, ylabel=None)\n",
    "ax = sns.set_style(\"white\")\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)\n",
    "plt.savefig(\"{}downstream_Vermeer.png\".format(output_path),  bbox_inches = 'tight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Could not interpret input '9'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-59-303d3a78ff62>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcatplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"9\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"model\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0medgecolor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\".4\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mresults\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkind\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"bar\"\u001b[0m\u001b[0;34m,\u001b[0m  \u001b[0mheight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m8\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_titles\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"{col_name} {col_var}\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdespine\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mleft\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxlabel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mylabel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_style\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"white\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlegend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbbox_to_anchor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1.05\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mloc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mborderaxespad\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/seaborn/categorical.py\u001b[0m in \u001b[0;36mcatplot\u001b[0;34m(x, y, hue, data, row, col, col_wrap, estimator, ci, n_boot, units, order, hue_order, row_order, col_order, kind, height, aspect, orient, color, palette, legend, legend_out, sharex, sharey, margin_titles, facet_kws, **kwargs)\u001b[0m\n\u001b[1;32m   3714\u001b[0m     \u001b[0;31m# facets to ensure representation of all data in the final plot\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3715\u001b[0m     \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_CategoricalPlotter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3716\u001b[0;31m     \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mestablish_variables\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morient\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue_order\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3717\u001b[0m     \u001b[0morder\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgroup_names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3718\u001b[0m     \u001b[0mhue_order\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhue_names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/seaborn/categorical.py\u001b[0m in \u001b[0;36mestablish_variables\u001b[0;34m(self, x, y, hue, data, orient, order, hue_order, units)\u001b[0m\n\u001b[1;32m    153\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstring_types\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    154\u001b[0m                     \u001b[0merr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"Could not interpret input '{}'\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 155\u001b[0;31m                     \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0merr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    156\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    157\u001b[0m             \u001b[0;31m# Figure out the plotting orientation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Could not interpret input '9'"
     ]
    }
   ],
   "source": [
    "ax = sns.catplot(x=\"9\", y=\"model\", edgecolor=\".4\", data=results, kind=\"bar\",  height=8)\n",
    "ax.set_titles(\"{col_name} {col_var}\").despine(left=True)\n",
    "ax.set(xlabel=None, ylabel=None)\n",
    "ax = sns.set_style(\"white\")\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "ename": "ValueError",
     "evalue": "Could not interpret input 'model'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-52-d6aa2ca6ad5a>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcatplot\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"macro avg\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"model\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0medgecolor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\".4\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mresults\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkind\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"bar\"\u001b[0m\u001b[0;34m,\u001b[0m  \u001b[0mheight\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m8\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_titles\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"{col_name} {col_var}\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdespine\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mleft\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0max\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mxlabel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mylabel\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mNone\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0msns\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mset_style\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"white\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlegend\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mbbox_to_anchor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1.05\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mloc\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mborderaxespad\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m0.\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/seaborn/categorical.py\u001b[0m in \u001b[0;36mcatplot\u001b[0;34m(x, y, hue, data, row, col, col_wrap, estimator, ci, n_boot, units, order, hue_order, row_order, col_order, kind, height, aspect, orient, color, palette, legend, legend_out, sharex, sharey, margin_titles, facet_kws, **kwargs)\u001b[0m\n\u001b[1;32m   3714\u001b[0m     \u001b[0;31m# facets to ensure representation of all data in the final plot\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3715\u001b[0m     \u001b[0mp\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_CategoricalPlotter\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 3716\u001b[0;31m     \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mestablish_variables\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdata\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morient\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0morder\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mhue_order\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   3717\u001b[0m     \u001b[0morder\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgroup_names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   3718\u001b[0m     \u001b[0mhue_order\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mp\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhue_names\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m/usr/local/lib/python3.5/dist-packages/seaborn/categorical.py\u001b[0m in \u001b[0;36mestablish_variables\u001b[0;34m(self, x, y, hue, data, orient, order, hue_order, units)\u001b[0m\n\u001b[1;32m    153\u001b[0m                 \u001b[0;32mif\u001b[0m \u001b[0misinstance\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstring_types\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    154\u001b[0m                     \u001b[0merr\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"Could not interpret input '{}'\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 155\u001b[0;31m                     \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0merr\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    156\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    157\u001b[0m             \u001b[0;31m# Figure out the plotting orientation\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mValueError\u001b[0m: Could not interpret input 'model'"
     ]
    }
   ],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No handles with labels found to put in legend.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x7f50307cc2b0>"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f50380cc390>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = sns.catplot(x=\"9\", y=\"model\", edgecolor=\".4\", data=results, kind=\"bar\",  height=8)\n",
    "ax.set_titles(\"{col_name} {col_var}\").despine(left=True)\n",
    "ax.set(xlabel=None, ylabel=None)\n",
    "ax = sns.set_style(\"white\")\n",
    "plt.legend(bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0.)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/home/anne/tmpanne/AEM_output/embeddings_true_predicted_dataset_vermeer_embed_size_large.json\n"
     ]
    },
    {
     "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>actual</th>\n",
       "      <th>classifier</th>\n",
       "      <th>model</th>\n",
       "      <th>predicted</th>\n",
       "      <th>vectorizer</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>wiki.nl.vec</td>\n",
       "      <td>[3, 1, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>wiki.nl.vec</td>\n",
       "      <td>[3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_5_window_10_size_300_negsample_5</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 1, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 3, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 2, 1, 3, 2, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>cow-320.txt</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_5_window_10_size_300_negsample_5</td>\n",
       "      <td>[3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 2, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_6_window_10_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_5_window_10_size_300_negsample_5</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_2_window_5_size_100_negsample_15</td>\n",
       "      <td>[3, 1, 1, 3, 1, 3, 1, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_2_window_5_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_2_window_5_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_2_window_5_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 3, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_2_window_5_size_100_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 4, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_3_window_5_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>wiki.nl.vec</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 2, 3, 1, 3, 1, 3, 1, 3, 1, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 4, 1, 3, 1, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_10_window_47615_size_100_negsampl...</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>wiki.nl.vec</td>\n",
       "      <td>[3, 3, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_count</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>BernoulliNB</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>[3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...</td>\n",
       "      <td>ET</td>\n",
       "      <td>w2v_model_nr_7_window_10_size_300_negsample_15</td>\n",
       "      <td>[3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>140 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                actual         classifier  \\\n",
       "0    [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "1    [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "10   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "100  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "101  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "102  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "103  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "104  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "105  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "106  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "107  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "108  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "109  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "11   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "110  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "111  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "112  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "113  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "114  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "115  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "116  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "117  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "118  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "119  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "12   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "120  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "121  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "122  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "123  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "124  [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "..                                                 ...                ...   \n",
       "72   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "73   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "74   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "75   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "76   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "77   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "78   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "79   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "8    [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "80   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "81   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "82   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "83   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "84   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "85   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "86   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "87   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "88   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "89   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "9    [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "90   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "91   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "92   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "93   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "94   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "95   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...        BernoulliNB   \n",
       "96   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...  Passive Agressive   \n",
       "97   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...      SGDClassifier   \n",
       "98   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                SVM   \n",
       "99   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 4, 3, 3, 3, ...                 ET   \n",
       "\n",
       "                                                 model  \\\n",
       "0                                          wiki.nl.vec   \n",
       "1                                          wiki.nl.vec   \n",
       "10       w2v_model_nr_5_window_10_size_300_negsample_5   \n",
       "100                                        cow-320.txt   \n",
       "101                                        cow-320.txt   \n",
       "102                                        cow-320.txt   \n",
       "103                                        cow-320.txt   \n",
       "104                                        cow-320.txt   \n",
       "105                                        cow-320.txt   \n",
       "106                                        cow-320.txt   \n",
       "107                                        cow-320.txt   \n",
       "108                                        cow-320.txt   \n",
       "109                                        cow-320.txt   \n",
       "11       w2v_model_nr_5_window_10_size_300_negsample_5   \n",
       "110     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "111     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "112     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "113     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "114     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "115     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "116     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "117     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "118     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "119     w2v_model_nr_6_window_10_size_100_negsample_15   \n",
       "12       w2v_model_nr_5_window_10_size_300_negsample_5   \n",
       "120      w2v_model_nr_2_window_5_size_100_negsample_15   \n",
       "121      w2v_model_nr_2_window_5_size_100_negsample_15   \n",
       "122      w2v_model_nr_2_window_5_size_100_negsample_15   \n",
       "123      w2v_model_nr_2_window_5_size_100_negsample_15   \n",
       "124      w2v_model_nr_2_window_5_size_100_negsample_15   \n",
       "..                                                 ...   \n",
       "72       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "73       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "74       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "75       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "76       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "77       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "78       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "79       w2v_model_nr_3_window_5_size_300_negsample_15   \n",
       "8                                          wiki.nl.vec   \n",
       "80   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "81   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "82   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "83   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "84   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "85   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "86   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "87   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "88   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "89   w2v_model_nr_10_window_47615_size_100_negsampl...   \n",
       "9                                          wiki.nl.vec   \n",
       "90      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "91      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "92      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "93      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "94      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "95      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "96      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "97      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "98      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "99      w2v_model_nr_7_window_10_size_300_negsample_15   \n",
       "\n",
       "                                             predicted vectorizer  \n",
       "0    [3, 1, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "1    [3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "10   [3, 2, 1, 3, 1, 3, 1, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "100  [3, 2, 3, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "101  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "102  [3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "103  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "104  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "105  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "106  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "107  [3, 2, 1, 2, 1, 3, 2, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "108  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "109  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "11   [3, 2, 1, 2, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "110  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 2, 1, 3, 3, 3, ...  w2v_count  \n",
       "111  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "112  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "113  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "114  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "115  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "116  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "117  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "118  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "119  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "12   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "120  [3, 1, 1, 3, 1, 3, 1, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "121  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "122  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "123  [3, 2, 1, 3, 1, 3, 3, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "124  [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "..                                                 ...        ...  \n",
       "72   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "73   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "74   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "75   [3, 2, 1, 4, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "76   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "77   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "78   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "79   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "8    [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "80   [3, 2, 2, 3, 1, 3, 1, 3, 1, 3, 1, 2, 3, 3, 3, ...  w2v_count  \n",
       "81   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "82   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "83   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "84   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "85   [3, 2, 1, 4, 1, 3, 1, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "86   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "87   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "88   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "89   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "9    [3, 3, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "90   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "91   [3, 2, 1, 1, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "92   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "93   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_count  \n",
       "94   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_count  \n",
       "95   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "96   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "97   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "98   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 2, 3, 3, 3, ...  w2v_tfidf  \n",
       "99   [3, 2, 1, 3, 1, 3, 4, 3, 1, 3, 3, 1, 3, 3, 3, ...  w2v_tfidf  \n",
       "\n",
       "[140 rows x 5 columns]"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pathtodata='/home/anne/tmpanne/AEM_output'\n",
    "dataset = 'vermeer'\n",
    "classifier = 'SGDClassifier'\n",
    "\n",
    "base = \"{}/embeddings_true_predicted_dataset_{}_embed_size_large.json\".format(pathtodata, dataset)\n",
    "print(base)\n",
    "df = pd.read_json(base)\n",
    "#df = df[df['Classifier'] == classifier]\n",
    "#return df\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_json('/home/anne/tmpanne/AEM_output/embeddings_true_predicted_dataset_vermeer_embed_size_large.json')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "class get_heatmaps():\n",
    "    '''Get those heatmaps'''\n",
    "\n",
    "    def __init__(self, approach, sample, classifier = None, vect=None):\n",
    "        self.approach = approach\n",
    "        self.sample = sample\n",
    "        self.classifier = classifier\n",
    "        self.vect = vect\n",
    "\n",
    "    def get_data(self):\n",
    "\n",
    "        base = \"{}/embeddings_true_predicted_dataset_{}_embed_size_large.json\".format(pathtodata, dataset)\n",
    "        df = pd.read_json(base)\n",
    "        df = df[df['classifier'] == self.classifier]\n",
    "        return df\n",
    "\n",
    "    def confusion_matrix(self):\n",
    "        df = self.get_data()\n",
    "        df.replace(self.translator, inplace=True)\n",
    "        df.replace(d2, inplace=True)\n",
    "        confusion_matrix = pd.crosstab(df['Actual label'], df['Predicted label'], rownames=['True'], colnames=['Predicted'])\n",
    "        cmn = confusion_matrix.astype('float') / confusion_matrix.sum(axis=1)[:, np.newaxis]\n",
    "        return cmn\n",
    "\n",
    "    def get_heatmap(self):\n",
    "        label = \"Values in the diagonal represent the relative times that the manual coding (’true label’ - Y axis ) is equal to the classifier (X axis). Diagonal values indicate the relative number of correct predictions: The higher the values in diagonal, the better the prediction. Off-diagonal values indicate misclassification. Darker colours indicate higher values. Due to class imbalance, values are normalised to facilitate visual understanding. Values below 0.1 are not visualised. \"\n",
    "        cmn = self.confusion_matrix()\n",
    "        cmn = cmn.round(1)\n",
    "        fig, ax = plt.subplots(figsize=(10,10))\n",
    "        heatmap = sns.heatmap(cmn, annot=True, annot_kws={\"size\": 10}, fmt='.1f',  cmap=\"BuGn\", mask=(cmn<0.1))\n",
    "    #    heatmap = sns.heatmap(cmn, annot=True, annot_kws={\"size\": 10}, fmt='.1f',  cmap=\"gist_gray_r\", linecolor='black', mask=(cmn<0.1))\n",
    "        for _, spine in heatmap.spines.items():\n",
    "            spine.set_visible(True)\n",
    "        fs = 16\n",
    "        heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels(), rotation=0, ha='right', fontsize=fs)\n",
    "        heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels(), rotation=0, ha='right', fontsize=fs)\n",
    "        #plt.title(\"Normalized Confusion Matrix (Classifier: {}) \".format(self.approach), fontsize= 16)\n",
    "        plt.title(None)\n",
    "        plt.ylabel('True label (Manual coding)', fontsize=fs)\n",
    "        plt.xlabel('Predicted label', fontsize=fs)\n",
    "        return fig\n",
    "\n",
    "    def get_figure_save(self):\n",
    "        #logger.info('{}'.format(label))\n",
    "        figure = self.get_heatmap()\n",
    "        if self.approach == 'SML':\n",
    "            fname = '{}Heatmap_{}_{}_{}'.format(OUTPUTFIGURES, self.approach, self.classifier, self.sample)\n",
    "        else:\n",
    "             fname = '{}Heatmap_{}_{}'.format(OUTPUTFIGURES, self.approach, self.sample)\n",
    "        figure.savefig(fname, bbox_inches='tight')\n",
    "        print('Saved figure as: {}'.format(fname))\n",
    "        \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "pathtodata='/home/anne/tmpanne/AEM_output'\n",
    "dataset = 'vermeer'\n",
    "classifier = 'SGDClassifier'\n",
    "\n",
    "\n",
    "a = get_heatmaps(classifier='SGDClassifier', vect='w2v_count', approach = 'SML', sample='Vermeer')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "ename": "AttributeError",
     "evalue": "'get_heatmaps' object has no attribute 'translator'",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-77-1037c3d79865>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0ma\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_heatmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[0;32m<ipython-input-75-6565374f3b51>\u001b[0m in \u001b[0;36mget_heatmap\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     27\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mget_heatmap\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     28\u001b[0m         \u001b[0mlabel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m\"Values in the diagonal represent the relative times that the manual coding (’true label’ - Y axis ) is equal to the classifier (X axis). Diagonal values indicate the relative number of correct predictions: The higher the values in diagonal, the better the prediction. Off-diagonal values indicate misclassification. Darker colours indicate higher values. Due to class imbalance, values are normalised to facilitate visual understanding. Values below 0.1 are not visualised. \"\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 29\u001b[0;31m         \u001b[0mcmn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mconfusion_matrix\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     30\u001b[0m         \u001b[0mcmn\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcmn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mround\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     31\u001b[0m         \u001b[0mfig\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0max\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mplt\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msubplots\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mfigsize\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m10\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;32m<ipython-input-75-6565374f3b51>\u001b[0m in \u001b[0;36mconfusion_matrix\u001b[0;34m(self)\u001b[0m\n\u001b[1;32m     19\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mconfusion_matrix\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     20\u001b[0m         \u001b[0mdf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mget_data\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 21\u001b[0;31m         \u001b[0mdf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtranslator\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minplace\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     22\u001b[0m         \u001b[0mdf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreplace\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0md2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minplace\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;32mTrue\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     23\u001b[0m         \u001b[0mconfusion_matrix\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcrosstab\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Actual label'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Predicted label'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mrownames\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'True'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcolnames\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'Predicted'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mAttributeError\u001b[0m: 'get_heatmaps' object has no attribute 'translator'"
     ]
    }
   ],
   "source": [
    "a.get_heatmap()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'v' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[0;32m<ipython-input-17-6b49cd042ca4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# select metric of interest\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mdf\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mv\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mv\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'metrics'\u001b[0m\u001b[0;34m]\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'f1-score'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0maveraging\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m'micro avg'\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0mresults\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0mdf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mgroupby\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'model'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maveraging\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransform\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmax\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0mdf\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0maveraging\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msort_values\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0maveraging\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
      "\u001b[0;31mNameError\u001b[0m: name 'v' is not defined"
     ]
    }
   ],
   "source": [
    "# select metric of interest\n",
    "df = v[v['metrics'] == 'f1-score']\n",
    "averaging = 'micro avg'\n",
    "\n",
    "results = df[df.groupby(['model'])[averaging].transform(max) == df[averaging]].sort_values(averaging)\n",
    "\n",
    "results[['classifier', 'vectorizer', averaging , 'clf__alpha', 'clf__fit_intercept', 'clf__gamma', 'clf__kernel',\n",
    "       'clf__loss', 'clf__max_features', 'clf__max_iter', 'clf__penalty']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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>model</th>\n",
       "      <th>vectorizer</th>\n",
       "      <th>metrics</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>macro avg</th>\n",
       "      <th>micro avg</th>\n",
       "      <th>...</th>\n",
       "      <th>clf__C</th>\n",
       "      <th>clf__alpha</th>\n",
       "      <th>clf__fit_intercept</th>\n",
       "      <th>clf__gamma</th>\n",
       "      <th>clf__kernel</th>\n",
       "      <th>clf__loss</th>\n",
       "      <th>clf__max_features</th>\n",
       "      <th>clf__max_iter</th>\n",
       "      <th>clf__penalty</th>\n",
       "      <th>indicators</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ET</td>\n",
       "      <td>baseline</td>\n",
       "      <td>Count</td>\n",
       "      <td>f1-score</td>\n",
       "      <td>0.700422</td>\n",
       "      <td>0.458333</td>\n",
       "      <td>0.816374</td>\n",
       "      <td>0.535714</td>\n",
       "      <td>0.627711</td>\n",
       "      <td>0.724928</td>\n",
       "      <td>...</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>auto</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>clf__max_features</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>ET</td>\n",
       "      <td>baseline</td>\n",
       "      <td>Tfidf</td>\n",
       "      <td>f1-score</td>\n",
       "      <td>0.735043</td>\n",
       "      <td>0.494737</td>\n",
       "      <td>0.815155</td>\n",
       "      <td>0.475248</td>\n",
       "      <td>0.630046</td>\n",
       "      <td>0.733524</td>\n",
       "      <td>...</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>auto</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>clf__max_features</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ET</td>\n",
       "      <td>baseline</td>\n",
       "      <td>Count</td>\n",
       "      <td>precision</td>\n",
       "      <td>0.674797</td>\n",
       "      <td>0.354839</td>\n",
       "      <td>0.930667</td>\n",
       "      <td>0.394737</td>\n",
       "      <td>0.588760</td>\n",
       "      <td>0.724928</td>\n",
       "      <td>...</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>auto</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>clf__max_features</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ET</td>\n",
       "      <td>baseline</td>\n",
       "      <td>Tfidf</td>\n",
       "      <td>precision</td>\n",
       "      <td>0.699187</td>\n",
       "      <td>0.379032</td>\n",
       "      <td>0.946667</td>\n",
       "      <td>0.315789</td>\n",
       "      <td>0.585169</td>\n",
       "      <td>0.733524</td>\n",
       "      <td>...</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>auto</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>clf__max_features</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ET</td>\n",
       "      <td>baseline</td>\n",
       "      <td>Count</td>\n",
       "      <td>recall</td>\n",
       "      <td>0.728070</td>\n",
       "      <td>0.647059</td>\n",
       "      <td>0.727083</td>\n",
       "      <td>0.833333</td>\n",
       "      <td>0.733886</td>\n",
       "      <td>0.724928</td>\n",
       "      <td>...</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>auto</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>clf__max_features</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  classifier     model vectorizer    metrics         1         2         3  \\\n",
       "0         ET  baseline      Count   f1-score  0.700422  0.458333  0.816374   \n",
       "1         ET  baseline      Tfidf   f1-score  0.735043  0.494737  0.815155   \n",
       "2         ET  baseline      Count  precision  0.674797  0.354839  0.930667   \n",
       "3         ET  baseline      Tfidf  precision  0.699187  0.379032  0.946667   \n",
       "4         ET  baseline      Count     recall  0.728070  0.647059  0.727083   \n",
       "\n",
       "          4  macro avg  micro avg  ...  clf__C  clf__alpha clf__fit_intercept  \\\n",
       "0  0.535714   0.627711   0.724928  ...     NaN         NaN                NaN   \n",
       "1  0.475248   0.630046   0.733524  ...     NaN         NaN                NaN   \n",
       "2  0.394737   0.588760   0.724928  ...     NaN         NaN                NaN   \n",
       "3  0.315789   0.585169   0.733524  ...     NaN         NaN                NaN   \n",
       "4  0.833333   0.733886   0.724928  ...     NaN         NaN                NaN   \n",
       "\n",
       "  clf__gamma clf__kernel clf__loss clf__max_features clf__max_iter  \\\n",
       "0        NaN         NaN       NaN              auto           NaN   \n",
       "1        NaN         NaN       NaN              auto           NaN   \n",
       "2        NaN         NaN       NaN              auto           NaN   \n",
       "3        NaN         NaN       NaN              auto           NaN   \n",
       "4        NaN         NaN       NaN              auto           NaN   \n",
       "\n",
       "  clf__penalty         indicators  \n",
       "0          NaN  clf__max_features  \n",
       "1          NaN  clf__max_features  \n",
       "2          NaN  clf__max_features  \n",
       "3          NaN  clf__max_features  \n",
       "4          NaN  clf__max_features  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()\n",
    "df.set_index('model', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "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>vectorizer</th>\n",
       "      <th>weighted avg</th>\n",
       "      <th>clf__alpha</th>\n",
       "      <th>clf__fit_intercept</th>\n",
       "      <th>clf__gamma</th>\n",
       "      <th>clf__kernel</th>\n",
       "      <th>clf__loss</th>\n",
       "      <th>clf__max_features</th>\n",
       "      <th>clf__max_iter</th>\n",
       "      <th>clf__penalty</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>model</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_10_window_47615_size_100_negsample_15</th>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_count</td>\n",
       "      <td>0.817128</td>\n",
       "      <td>0.01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>elasticnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_2_window_5_size_100_negsample_15</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.817608</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>rbf</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>baseline</th>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>Tfidf</td>\n",
       "      <td>0.819430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>hinge</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_8_window_47615_size_100_negsample_5</th>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.821569</td>\n",
       "      <td>0.01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>l2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>wiki.nl.vec</th>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_count</td>\n",
       "      <td>0.823054</td>\n",
       "      <td>1e-05</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>elasticnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cow-320.txt</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.826648</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>linear</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_1_window_5_size_300_negsample_5</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.826838</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>rbf</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_3_window_5_size_300_negsample_15</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.826907</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>rbf</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_4_window_10_size_100_negsample_5</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.826925</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>rbf</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_0_window_5_size_100_negsample_5</th>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.828821</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>hinge</td>\n",
       "      <td>NaN</td>\n",
       "      <td>10</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_7_window_10_size_300_negsample_15</th>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.829659</td>\n",
       "      <td>0.01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>l2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_6_window_10_size_100_negsample_15</th>\n",
       "      <td>SGDClassifier</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.831459</td>\n",
       "      <td>0.01</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>30</td>\n",
       "      <td>elasticnet</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_11_window_47615_size_300_negsample_15</th>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.832206</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>squared_hinge</td>\n",
       "      <td>NaN</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_5_window_10_size_300_negsample_5</th>\n",
       "      <td>SVM</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.832312</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.001</td>\n",
       "      <td>rbf</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>w2v_model_nr_9_window_47615_size_300_negsample_5</th>\n",
       "      <td>Passive Agressive</td>\n",
       "      <td>w2v_tfidf</td>\n",
       "      <td>0.835331</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>hinge</td>\n",
       "      <td>NaN</td>\n",
       "      <td>15</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                                           classifier  \\\n",
       "model                                                                   \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15      SGDClassifier   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15                     SVM   \n",
       "baseline                                            Passive Agressive   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5        SGDClassifier   \n",
       "wiki.nl.vec                                             SGDClassifier   \n",
       "cow-320.txt                                                       SVM   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5                      SVM   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15                     SVM   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5                     SVM   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5        Passive Agressive   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15          SGDClassifier   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15          SGDClassifier   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15  Passive Agressive   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5                     SVM   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5    Passive Agressive   \n",
       "\n",
       "                                                   vectorizer  weighted avg  \\\n",
       "model                                                                         \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15  w2v_count      0.817128   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15       w2v_tfidf      0.817608   \n",
       "baseline                                                Tfidf      0.819430   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5    w2v_tfidf      0.821569   \n",
       "wiki.nl.vec                                         w2v_count      0.823054   \n",
       "cow-320.txt                                         w2v_tfidf      0.826648   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5        w2v_tfidf      0.826838   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15       w2v_tfidf      0.826907   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5       w2v_tfidf      0.826925   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5        w2v_tfidf      0.828821   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15      w2v_tfidf      0.829659   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15      w2v_tfidf      0.831459   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15  w2v_tfidf      0.832206   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5       w2v_tfidf      0.832312   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5    w2v_tfidf      0.835331   \n",
       "\n",
       "                                                   clf__alpha  \\\n",
       "model                                                           \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15       0.01   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15             NaN   \n",
       "baseline                                                  NaN   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5         0.01   \n",
       "wiki.nl.vec                                             1e-05   \n",
       "cow-320.txt                                               NaN   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5              NaN   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15             NaN   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5             NaN   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5              NaN   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15           0.01   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15           0.01   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15        NaN   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5             NaN   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5          NaN   \n",
       "\n",
       "                                                   clf__fit_intercept  \\\n",
       "model                                                                   \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15                NaN   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15                     NaN   \n",
       "baseline                                                         True   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5                  NaN   \n",
       "wiki.nl.vec                                                       NaN   \n",
       "cow-320.txt                                                       NaN   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5                      NaN   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15                     NaN   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5                     NaN   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5                    False   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15                    NaN   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15                    NaN   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15              False   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5                     NaN   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5                False   \n",
       "\n",
       "                                                   clf__gamma clf__kernel  \\\n",
       "model                                                                       \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15        NaN         NaN   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15           0.001         rbf   \n",
       "baseline                                                  NaN         NaN   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5          NaN         NaN   \n",
       "wiki.nl.vec                                               NaN         NaN   \n",
       "cow-320.txt                                             0.001      linear   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5            0.001         rbf   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15           0.001         rbf   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5           0.001         rbf   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5              NaN         NaN   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15            NaN         NaN   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15            NaN         NaN   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15        NaN         NaN   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5           0.001         rbf   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5          NaN         NaN   \n",
       "\n",
       "                                                        clf__loss  \\\n",
       "model                                                               \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15            NaN   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15                 NaN   \n",
       "baseline                                                    hinge   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5              NaN   \n",
       "wiki.nl.vec                                                   NaN   \n",
       "cow-320.txt                                                   NaN   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5                  NaN   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15                 NaN   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5                 NaN   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5                hinge   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15                NaN   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15                NaN   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15  squared_hinge   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5                 NaN   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5            hinge   \n",
       "\n",
       "                                                   clf__max_features  \\\n",
       "model                                                                  \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15               NaN   \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15                    NaN   \n",
       "baseline                                                         NaN   \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5                 NaN   \n",
       "wiki.nl.vec                                                      NaN   \n",
       "cow-320.txt                                                      NaN   \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5                     NaN   \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15                    NaN   \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5                    NaN   \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5                     NaN   \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15                   NaN   \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15                   NaN   \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15               NaN   \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5                    NaN   \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5                 NaN   \n",
       "\n",
       "                                                   clf__max_iter clf__penalty  \n",
       "model                                                                          \n",
       "w2v_model_nr_10_window_47615_size_100_negsample_15            30   elasticnet  \n",
       "w2v_model_nr_2_window_5_size_100_negsample_15                NaN          NaN  \n",
       "baseline                                                      15          NaN  \n",
       "w2v_model_nr_8_window_47615_size_100_negsample_5              30           l2  \n",
       "wiki.nl.vec                                                   30   elasticnet  \n",
       "cow-320.txt                                                  NaN          NaN  \n",
       "w2v_model_nr_1_window_5_size_300_negsample_5                 NaN          NaN  \n",
       "w2v_model_nr_3_window_5_size_300_negsample_15                NaN          NaN  \n",
       "w2v_model_nr_4_window_10_size_100_negsample_5                NaN          NaN  \n",
       "w2v_model_nr_0_window_5_size_100_negsample_5                  10          NaN  \n",
       "w2v_model_nr_7_window_10_size_300_negsample_15                30           l2  \n",
       "w2v_model_nr_6_window_10_size_100_negsample_15                30   elasticnet  \n",
       "w2v_model_nr_11_window_47615_size_300_negsample_15             5          NaN  \n",
       "w2v_model_nr_5_window_10_size_300_negsample_5                NaN          NaN  \n",
       "w2v_model_nr_9_window_47615_size_300_negsample_5              15          NaN  "
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# select metric of interest\n",
    "df = df[df['metrics'] == 'f1-score']\n",
    "\n",
    "results = df[df.groupby(['model'])['weighted avg'].transform(max) == df['weighted avg']].sort_values('weighted avg')\n",
    "results[['classifier', 'vectorizer', 'weighted avg', 'clf__alpha', 'clf__fit_intercept', 'clf__gamma', 'clf__kernel',\n",
    "       'clf__loss', 'clf__max_features', 'clf__max_iter', 'clf__penalty']]"
   ]
  }
 ],
 "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": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}