comparison notebooks/sensitivity_experiment_outliers.ipynb @ 94:69521f86d931 branch-tests

notebooks for publication

+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "The autoreload extension is already loaded. To reload it, use:\n",
+      "  %reload_ext autoreload\n"
+     ]
+    }
+   ],
+   "source": [
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from sklearn.model_selection import train_test_split\n",
+    "from collections import Counter\n",
+    "\n",
+    "%matplotlib inline\n",
+    "import matplotlib.pyplot as plt\n",
+    "\n",
+    "%load_ext autoreload\n",
+    "%autoreload 2\n",
+    "\n",
+    "import sys\n",
+    "sys.path.append('../')\n",
+    "import scripts.classification as classification\n",
+    "import scripts.outliers as outliers"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Sample 80% of the dataset, for 10 times"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's sample only 80% of the recordings each time (in a stratified manner) so that the set of recordings considered for each country is changed every time."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": [
+    "results_file = '../data/lda_data_8.pickle'\n",
+    "n_iters = 10\n",
+    "for n in range(n_iters):\n",
+    "    print \"iteration %d\" % n\n",
+    "    print results_file\n",
+    "    X, Y, Yaudio = classification.load_data_from_pickle(results_file)\n",
+    "    # get only 80% of the dataset.. to vary the choice of outliers\n",
+    "    X, _, Y, _ = train_test_split(X, Y, train_size=0.8, stratify=Y)\n",
+    "    print X.shape, Y.shape\n",
+    "    # outliers\n",
+    "    print \"detecting outliers...\"\n",
+    "    df_global, threshold, MD = outliers.get_outliers_df(X, Y, chi2thr=0.999)\n",
+    "    outliers.print_most_least_outliers_topN(df_global, N=10)\n",
+    "    \n",
+    "    # write output\n",
+    "    print \"writing file\"\n",
+    "    df_global.to_csv('../data/outliers_'+str(n)+'.csv', index=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "n_iters = 10\n",
+    "ranked_countries = pd.DataFrame()\n",
+    "ranked_outliers = pd.DataFrame()\n",
+    "for n in range(n_iters):\n",
+    "    df_global = pd.read_csv('../data/outliers_'+str(n)+'.csv')\n",
+    "    df_global = df_global.sort_values('Outliers', axis=0, ascending=False).reset_index()\n",
+    "    ranked_countries = pd.concat([ranked_countries, df_global['Country']], axis=1)\n",
+    "    ranked_outliers = pd.concat([ranked_outliers, df_global['Outliers']], axis=1)\n",
+    "ranked_countries_arr = ranked_countries.get_values()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Estimate precision at K"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "First get the ground truth from a majority vote on the top K=10 positions."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": [
+    "# majority voting + precision at K\n",
+    "K_vote = 10\n",
+    "country_vote = Counter(ranked_countries_arr[:K_vote, :].ravel())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>index</th>\n",
+       "      <th>0</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Pakistan</td>\n",
+       "      <td>10</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Chad</td>\n",
+       "      <td>10</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>5</th>\n",
+       "      <td>Gambia</td>\n",
+       "      <td>10</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>10</th>\n",
+       "      <td>Ivory Coast</td>\n",
+       "      <td>10</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>12</th>\n",
+       "      <td>Botswana</td>\n",
+       "      <td>10</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>6</th>\n",
+       "      <td>Nepal</td>\n",
+       "      <td>9</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>13</th>\n",
+       "      <td>Benin</td>\n",
+       "      <td>8</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8</th>\n",
+       "      <td>Senegal</td>\n",
+       "      <td>7</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>9</th>\n",
+       "      <td>French Guiana</td>\n",
+       "      <td>7</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>El Salvador</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>11</th>\n",
+       "      <td>Mozambique</td>\n",
+       "      <td>5</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>Uganda</td>\n",
+       "      <td>4</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Bhutan</td>\n",
+       "      <td>3</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Liberia</td>\n",
+       "      <td>2</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "            index   0\n",
+       "0        Pakistan  10\n",
+       "2            Chad  10\n",
+       "5          Gambia  10\n",
+       "10    Ivory Coast  10\n",
+       "12       Botswana  10\n",
+       "6           Nepal   9\n",
+       "13          Benin   8\n",
+       "8         Senegal   7\n",
+       "9   French Guiana   7\n",
+       "4     El Salvador   5\n",
+       "11     Mozambique   5\n",
+       "7          Uganda   4\n",
+       "1          Bhutan   3\n",
+       "3         Liberia   2"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_country_vote = pd.DataFrame.from_dict(country_vote, orient='index').reset_index()\n",
+    "df_country_vote.sort_values(0, ascending=False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": [
+    "def precision_at_k(array, gr_truth, k):\n",
+    "    return len(set(array[:k]) & set(gr_truth[:k])) / float(k)\n",
+    "    \n",
+    "k = 10\n",
+    "ground_truth = df_country_vote['index'].get_values()\n",
+    "p_ = []\n",
+    "for j in range(ranked_countries_arr.shape[1]):\n",
+    "    p_.append(precision_at_k(ranked_countries_arr[:, j], ground_truth, k))\n",
+    "p_ = np.array(p_)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "mean 0.67\n",
+      "std 0.0640312423743\n"
+     ]
+    }
+   ],
+   "source": [
+    "print 'mean', np.mean(p_) \n",
+    "print 'std', np.std(p_)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "[ 0.6  0.7  0.7  0.6  0.6  0.7  0.8  0.6  0.7  0.7]\n"
+     ]
+    }
+   ],
+   "source": [
+    "print p_"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {
+    "collapsed": true
+   },
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 2",
+   "language": "python",
+   "name": "python2"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 2
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython2",
+   "version": "2.7.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}