FS_DR.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Feature selection on detector regions\n",
    "We consider detector regions to be all combinations of volumes with their associated layers. When we look at the detector layout, any horizontal or vertical bar is considered to be a detector region.\n",
    "\n",
    "<img src=\"images/Detector.png\" alt=\"Detector layout\" width=\"600\"/>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from scripts.tracks import get_tracks\n",
    "from sklearn.metrics import roc_auc_score\n",
    "from sklearn.preprocessing import label_binarize\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Unpickle FE dataframe and FS feature importances"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(120939, 46)\n",
      "(38, 5)\n"
     ]
    }
   ],
   "source": [
    "event_id = 'event000001000'\n",
    "data = pd.read_pickle('/home/alepfu/Desktop/dataset/pickle/' + event_id + '_FE.pkl')\n",
    "print(data.shape)\n",
    "\n",
    "feature_importances = pd.read_pickle('/home/alepfu/Desktop/dataset/pickle/feature_importances.pkl')\n",
    "print(feature_importances.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Pick a sample of tracks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(120939, 46)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "N_tracks = 0\n",
    "data = get_tracks(data, N_tracks=N_tracks, drop_zero_weights=False, drop_zero_id=False, factorize_particle_id=False)\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Pick a subset of features\n",
    "For now we only keep high ranking features (high mean importance), see FS notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['x_r_sph', 'y_r_sph', 'cos_dist_rc', 'phi', 'cos_dist_lc',\n",
       "       'z_r_cyl', 'z_r_sph', 'theta', 'phi_atan2', 'y_r_cyl',\n",
       "       'theta_atan2', 'x_r_cyl', 'theta_pixel'], dtype=object)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "N_features = 13\n",
    "features = feature_importances.head(N_features)['feature'].values\n",
    "features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(120939, 16)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data = data[np.insert(features, 0, ['hit_id','particle_id','vol_lay_concat'], axis=0)].copy()\n",
    "data.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Set detector regions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of detector regions: 7\n"
     ]
    }
   ],
   "source": [
    "regions = np.array([72,74,76,78,710,712,714])\n",
    "\n",
    "#\n",
    "#,82,84,86,88,92,94,96,98,910,912,914,\n",
    "#                    122,124,126,128,1210,1212,132,134,136,138,142,144,146,148,1410,1412,\n",
    "#                    162,164,166,168,1610,1612,172,174,182,184,186,188,1810,1812\n",
    "\n",
    "print('Number of detector regions:', len(regions))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Atm we keep it simple and only use decsion tree for feature importances, later we add other trees."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "CPU times: user 1min 53s, sys: 268 ms, total: 1min 54s\n",
      "Wall time: 1min 54s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "mm_scaler = MinMaxScaler(feature_range=(0,100))\n",
    "\n",
    "imps = []\n",
    "nhits = []\n",
    "scores = []\n",
    "roc_auc = []\n",
    "\n",
    "for reg in regions:\n",
    "    data_reg = data[data['vol_lay_concat'] == reg]\n",
    "    nhits.append(data_reg.shape[0])\n",
    "    \n",
    "    X = data_reg.iloc[:,3:]\n",
    "    y = data_reg[['particle_id']].values.ravel()\n",
    "    y = label_binarize(y, classes=list(np.unique(y)))\n",
    "    \n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)\n",
    "    \n",
    "    dec_tree = DecisionTreeClassifier(random_state=0)\n",
    "    dec_tree.fit(X_train, y_train)   \n",
    "    dec_tree_preds = dec_tree.predict(X_test)\n",
    "    roc_auc.append(roc_auc_score(y_test, dec_tree_preds, average='micro'))\n",
    "    scores.append(dec_tree.score(X_test, y_test))\n",
    "    \n",
    "    imps_scaled = mm_scaler.fit_transform(dec_tree.feature_importances_.reshape(-1,1))\n",
    "    imps_scaled = imps_scaled.reshape(1,-1).ravel()\n",
    "    imps.append(imps_scaled) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>region</th>\n",
       "      <th>x_r_sph</th>\n",
       "      <th>y_r_sph</th>\n",
       "      <th>cos_dist_rc</th>\n",
       "      <th>phi</th>\n",
       "      <th>cos_dist_lc</th>\n",
       "      <th>z_r_cyl</th>\n",
       "      <th>z_r_sph</th>\n",
       "      <th>theta</th>\n",
       "      <th>phi_atan2</th>\n",
       "      <th>y_r_cyl</th>\n",
       "      <th>theta_atan2</th>\n",
       "      <th>x_r_cyl</th>\n",
       "      <th>theta_pixel</th>\n",
       "      <th>nhits</th>\n",
       "      <th>score</th>\n",
       "      <th>roc_auc</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>72</td>\n",
       "      <td>73.177397</td>\n",
       "      <td>73.383202</td>\n",
       "      <td>11.788410</td>\n",
       "      <td>8.213213</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>27.851802</td>\n",
       "      <td>38.188315</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>31.230720</td>\n",
       "      <td>68.430182</td>\n",
       "      <td>58.206453</td>\n",
       "      <td>72.147596</td>\n",
       "      <td>20.979176</td>\n",
       "      <td>1449</td>\n",
       "      <td>0.379310</td>\n",
       "      <td>0.689338</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>74</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>76.927999</td>\n",
       "      <td>14.745343</td>\n",
       "      <td>4.902731</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>39.473624</td>\n",
       "      <td>11.627311</td>\n",
       "      <td>82.444196</td>\n",
       "      <td>45.781014</td>\n",
       "      <td>48.163314</td>\n",
       "      <td>68.726710</td>\n",
       "      <td>77.375568</td>\n",
       "      <td>33.263509</td>\n",
       "      <td>1767</td>\n",
       "      <td>0.511299</td>\n",
       "      <td>0.755436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>76</td>\n",
       "      <td>94.724051</td>\n",
       "      <td>80.687008</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>18.279984</td>\n",
       "      <td>19.689286</td>\n",
       "      <td>22.275643</td>\n",
       "      <td>10.767096</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>30.740733</td>\n",
       "      <td>46.374739</td>\n",
       "      <td>68.311839</td>\n",
       "      <td>38.152133</td>\n",
       "      <td>27.997853</td>\n",
       "      <td>2165</td>\n",
       "      <td>0.487298</td>\n",
       "      <td>0.743455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>78</td>\n",
       "      <td>76.150639</td>\n",
       "      <td>80.510404</td>\n",
       "      <td>15.133191</td>\n",
       "      <td>7.018348</td>\n",
       "      <td>19.729615</td>\n",
       "      <td>24.121597</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>17.982871</td>\n",
       "      <td>63.967178</td>\n",
       "      <td>89.339147</td>\n",
       "      <td>49.650137</td>\n",
       "      <td>12.861490</td>\n",
       "      <td>2553</td>\n",
       "      <td>0.547945</td>\n",
       "      <td>0.773819</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>710</td>\n",
       "      <td>97.033072</td>\n",
       "      <td>96.528735</td>\n",
       "      <td>18.463109</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.351617</td>\n",
       "      <td>20.706081</td>\n",
       "      <td>24.091733</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>2.268555</td>\n",
       "      <td>65.879075</td>\n",
       "      <td>88.579494</td>\n",
       "      <td>87.605492</td>\n",
       "      <td>16.089446</td>\n",
       "      <td>2887</td>\n",
       "      <td>0.605536</td>\n",
       "      <td>0.802647</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>712</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>78.223048</td>\n",
       "      <td>4.791410</td>\n",
       "      <td>21.476829</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.024627</td>\n",
       "      <td>2.319841</td>\n",
       "      <td>60.774871</td>\n",
       "      <td>3.692889</td>\n",
       "      <td>37.480382</td>\n",
       "      <td>88.569992</td>\n",
       "      <td>63.320367</td>\n",
       "      <td>2.312751</td>\n",
       "      <td>3028</td>\n",
       "      <td>0.594059</td>\n",
       "      <td>0.796908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>714</td>\n",
       "      <td>78.298026</td>\n",
       "      <td>57.893572</td>\n",
       "      <td>10.588295</td>\n",
       "      <td>27.085934</td>\n",
       "      <td>9.860668</td>\n",
       "      <td>9.052721</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>72.182646</td>\n",
       "      <td>7.551399</td>\n",
       "      <td>35.962661</td>\n",
       "      <td>100.000000</td>\n",
       "      <td>68.908976</td>\n",
       "      <td>8.694888</td>\n",
       "      <td>3024</td>\n",
       "      <td>0.631405</td>\n",
       "      <td>0.815594</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   region     x_r_sph    y_r_sph  cos_dist_rc        phi  cos_dist_lc  \\\n",
       "0      72   73.177397  73.383202    11.788410   8.213213     0.000000   \n",
       "1      74  100.000000  76.927999    14.745343   4.902731     0.000000   \n",
       "2      76   94.724051  80.687008     0.000000  18.279984    19.689286   \n",
       "3      78   76.150639  80.510404    15.133191   7.018348    19.729615   \n",
       "4     710   97.033072  96.528735    18.463109   0.000000     0.351617   \n",
       "5     712  100.000000  78.223048     4.791410  21.476829     0.000000   \n",
       "6     714   78.298026  57.893572    10.588295  27.085934     9.860668   \n",
       "\n",
       "     z_r_cyl    z_r_sph       theta  phi_atan2    y_r_cyl  theta_atan2  \\\n",
       "0  27.851802  38.188315  100.000000  31.230720  68.430182    58.206453   \n",
       "1  39.473624  11.627311   82.444196  45.781014  48.163314    68.726710   \n",
       "2  22.275643  10.767096  100.000000  30.740733  46.374739    68.311839   \n",
       "3  24.121597   0.000000  100.000000  17.982871  63.967178    89.339147   \n",
       "4  20.706081  24.091733  100.000000   2.268555  65.879075    88.579494   \n",
       "5   3.024627   2.319841   60.774871   3.692889  37.480382    88.569992   \n",
       "6   9.052721   0.000000   72.182646   7.551399  35.962661   100.000000   \n",
       "\n",
       "     x_r_cyl  theta_pixel  nhits     score   roc_auc  \n",
       "0  72.147596    20.979176   1449  0.379310  0.689338  \n",
       "1  77.375568    33.263509   1767  0.511299  0.755436  \n",
       "2  38.152133    27.997853   2165  0.487298  0.743455  \n",
       "3  49.650137    12.861490   2553  0.547945  0.773819  \n",
       "4  87.605492    16.089446   2887  0.605536  0.802647  \n",
       "5  63.320367     2.312751   3028  0.594059  0.796908  \n",
       "6  68.908976     8.694888   3024  0.631405  0.815594  "
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dec_tree_df = pd.concat([pd.DataFrame(regions, columns=['region']), \n",
    "                         pd.DataFrame(np.array(imps), columns=features), \n",
    "                         pd.DataFrame(nhits, columns=['nhits']), \n",
    "                         pd.DataFrame(scores, columns=['score']),\n",
    "                         pd.DataFrame(roc_auc, columns=['roc_auc'])], axis=1)\n",
    "\n",
    "dec_tree_df.head(len(regions))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Viz feature importances for every detector region"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fe8403ac5c0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = dec_tree_df[np.insert(features, 0, ['region'], axis=0)].plot(x='region', kind='bar', figsize=(30,10), colormap='tab20')\n",
    "ax.legend(loc='center left', bbox_to_anchor=(1, 0.5))\n",
    "ax.set_xticklabels(ax.get_xticklabels(), rotation=0);\n",
    "ax.set_xlabel('Detector region', fontsize=25);\n",
    "ax.set_ylabel('Importance',fontsize=25);\n",
    "ax.set_title('Decision tree feature importances for volume 7', fontsize=30);\n",
    "ax.tick_params(labelsize=20);\n",
    "plt.setp(ax.get_legend().get_texts(), fontsize='20');"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Isolate features that change a lot from layer to layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "x_r_sph        144.569891\n",
       "y_r_sph        130.521921\n",
       "cos_dist_rc     41.138213\n",
       "phi             98.182608\n",
       "cos_dist_lc     87.064764\n",
       "z_r_cyl        144.249444\n",
       "z_r_sph        202.353578\n",
       "theta          266.368693\n",
       "phi_atan2      274.239138\n",
       "y_r_cyl        186.574032\n",
       "theta_atan2    228.750276\n",
       "x_r_cyl        281.518650\n",
       "theta_pixel    117.078805\n",
       "dtype: float64"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dec_tree_df.iloc[:,1:-3].var(axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>x_r_sph</th>\n",
       "      <td>144.569891</td>\n",
       "      <td>144.569891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>y_r_sph</th>\n",
       "      <td>130.521921</td>\n",
       "      <td>130.521921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cos_dist_rc</th>\n",
       "      <td>41.138213</td>\n",
       "      <td>41.138213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>phi</th>\n",
       "      <td>98.182608</td>\n",
       "      <td>98.182608</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>cos_dist_lc</th>\n",
       "      <td>87.064764</td>\n",
       "      <td>87.064764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>z_r_cyl</th>\n",
       "      <td>144.249444</td>\n",
       "      <td>144.249444</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>z_r_sph</th>\n",
       "      <td>202.353578</td>\n",
       "      <td>202.353578</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>theta</th>\n",
       "      <td>266.368693</td>\n",
       "      <td>266.368693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>phi_atan2</th>\n",
       "      <td>274.239138</td>\n",
       "      <td>274.239138</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>y_r_cyl</th>\n",
       "      <td>186.574032</td>\n",
       "      <td>186.574032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>theta_atan2</th>\n",
       "      <td>228.750276</td>\n",
       "      <td>228.750276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>x_r_cyl</th>\n",
       "      <td>281.518650</td>\n",
       "      <td>281.518650</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>theta_pixel</th>\n",
       "      <td>117.078805</td>\n",
       "      <td>117.078805</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      0           1\n",
       "x_r_sph      144.569891  144.569891\n",
       "y_r_sph      130.521921  130.521921\n",
       "cos_dist_rc   41.138213   41.138213\n",
       "phi           98.182608   98.182608\n",
       "cos_dist_lc   87.064764   87.064764\n",
       "z_r_cyl      144.249444  144.249444\n",
       "z_r_sph      202.353578  202.353578\n",
       "theta        266.368693  266.368693\n",
       "phi_atan2    274.239138  274.239138\n",
       "y_r_cyl      186.574032  186.574032\n",
       "theta_atan2  228.750276  228.750276\n",
       "x_r_cyl      281.518650  281.518650\n",
       "theta_pixel  117.078805  117.078805"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "\n",
    "var_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}