Baseline - simple.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "import xgboost\n",
    "import utils\n",
    "import scoring\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA_PATH = \"./data\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/mnt/nkazeev/miniconda3/envs/pip-gpu/lib/python3.6/site-packages/numpy/lib/arraysetops.py:518: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  mask |= (ar1 == a)\n"
     ]
    }
   ],
   "source": [
    "train, test = utils.load_data_csv(\n",
    "    DATA_PATH, utils.SIMPLE_FEATURE_COLUMNS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>ncl[0]</th>\n",
       "      <th>ncl[1]</th>\n",
       "      <th>ncl[2]</th>\n",
       "      <th>ncl[3]</th>\n",
       "      <th>avg_cs[0]</th>\n",
       "      <th>avg_cs[1]</th>\n",
       "      <th>avg_cs[2]</th>\n",
       "      <th>avg_cs[3]</th>\n",
       "      <th>ndof</th>\n",
       "      <th>MatchedHit_TYPE[0]</th>\n",
       "      <th>...</th>\n",
       "      <th>Mextra_DX2[3]</th>\n",
       "      <th>Mextra_DY2[0]</th>\n",
       "      <th>Mextra_DY2[1]</th>\n",
       "      <th>Mextra_DY2[2]</th>\n",
       "      <th>Mextra_DY2[3]</th>\n",
       "      <th>FOI_hits_N</th>\n",
       "      <th>PT</th>\n",
       "      <th>P</th>\n",
       "      <th>label</th>\n",
       "      <th>weight</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>47</td>\n",
       "      <td>31</td>\n",
       "      <td>13</td>\n",
       "      <td>15</td>\n",
       "      <td>2.000000</td>\n",
       "      <td>1.580645</td>\n",
       "      <td>1.153846</td>\n",
       "      <td>1.133333</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>4295.08540</td>\n",
       "      <td>333.436300</td>\n",
       "      <td>951.68540</td>\n",
       "      <td>2125.09990</td>\n",
       "      <td>4145.18160</td>\n",
       "      <td>6</td>\n",
       "      <td>1764.925899</td>\n",
       "      <td>12646.817591</td>\n",
       "      <td>1</td>\n",
       "      <td>2.226325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>92</td>\n",
       "      <td>19</td>\n",
       "      <td>11</td>\n",
       "      <td>26</td>\n",
       "      <td>2.750000</td>\n",
       "      <td>2.789474</td>\n",
       "      <td>1.363636</td>\n",
       "      <td>1.230769</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>777.42914</td>\n",
       "      <td>64.267190</td>\n",
       "      <td>185.53355</td>\n",
       "      <td>413.95690</td>\n",
       "      <td>793.56604</td>\n",
       "      <td>4</td>\n",
       "      <td>4511.281266</td>\n",
       "      <td>28556.630027</td>\n",
       "      <td>1</td>\n",
       "      <td>2.127109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100</td>\n",
       "      <td>21</td>\n",
       "      <td>11</td>\n",
       "      <td>12</td>\n",
       "      <td>2.930000</td>\n",
       "      <td>2.428571</td>\n",
       "      <td>1.181818</td>\n",
       "      <td>1.083333</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>1931.48400</td>\n",
       "      <td>159.182330</td>\n",
       "      <td>440.46994</td>\n",
       "      <td>979.69520</td>\n",
       "      <td>1899.83920</td>\n",
       "      <td>9</td>\n",
       "      <td>1062.550086</td>\n",
       "      <td>17491.702634</td>\n",
       "      <td>1</td>\n",
       "      <td>4.305235</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>74</td>\n",
       "      <td>27</td>\n",
       "      <td>13</td>\n",
       "      <td>20</td>\n",
       "      <td>3.067568</td>\n",
       "      <td>2.518518</td>\n",
       "      <td>1.923077</td>\n",
       "      <td>1.300000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>443.21832</td>\n",
       "      <td>36.024845</td>\n",
       "      <td>103.41968</td>\n",
       "      <td>229.75415</td>\n",
       "      <td>438.86194</td>\n",
       "      <td>5</td>\n",
       "      <td>4991.074825</td>\n",
       "      <td>37192.868559</td>\n",
       "      <td>1</td>\n",
       "      <td>1.605583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>18</td>\n",
       "      <td>22</td>\n",
       "      <td>8</td>\n",
       "      <td>8</td>\n",
       "      <td>1.833333</td>\n",
       "      <td>1.954546</td>\n",
       "      <td>1.125000</td>\n",
       "      <td>1.750000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>2430.53470</td>\n",
       "      <td>190.557480</td>\n",
       "      <td>551.34094</td>\n",
       "      <td>1240.64840</td>\n",
       "      <td>2402.91920</td>\n",
       "      <td>6</td>\n",
       "      <td>1347.704888</td>\n",
       "      <td>16688.746174</td>\n",
       "      <td>1</td>\n",
       "      <td>2.055900</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 67 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   ncl[0]  ncl[1]  ncl[2]  ncl[3]  avg_cs[0]  avg_cs[1]  avg_cs[2]  avg_cs[3]  \\\n",
       "0      47      31      13      15   2.000000   1.580645   1.153846   1.133333   \n",
       "1      92      19      11      26   2.750000   2.789474   1.363636   1.230769   \n",
       "2     100      21      11      12   2.930000   2.428571   1.181818   1.083333   \n",
       "3      74      27      13      20   3.067568   2.518518   1.923077   1.300000   \n",
       "4      18      22       8       8   1.833333   1.954546   1.125000   1.750000   \n",
       "\n",
       "   ndof  MatchedHit_TYPE[0]    ...     Mextra_DX2[3]  Mextra_DY2[0]  \\\n",
       "0     8                   2    ...        4295.08540     333.436300   \n",
       "1     8                   2    ...         777.42914      64.267190   \n",
       "2     8                   2    ...        1931.48400     159.182330   \n",
       "3     8                   2    ...         443.21832      36.024845   \n",
       "4     8                   2    ...        2430.53470     190.557480   \n",
       "\n",
       "   Mextra_DY2[1]  Mextra_DY2[2]  Mextra_DY2[3]  FOI_hits_N           PT  \\\n",
       "0      951.68540     2125.09990     4145.18160           6  1764.925899   \n",
       "1      185.53355      413.95690      793.56604           4  4511.281266   \n",
       "2      440.46994      979.69520     1899.83920           9  1062.550086   \n",
       "3      103.41968      229.75415      438.86194           5  4991.074825   \n",
       "4      551.34094     1240.64840     2402.91920           6  1347.704888   \n",
       "\n",
       "              P  label    weight  \n",
       "0  12646.817591      1  2.226325  \n",
       "1  28556.630027      1  2.127109  \n",
       "2  17491.702634      1  4.305235  \n",
       "3  37192.868559      1  1.605583  \n",
       "4  16688.746174      1  2.055900  \n",
       "\n",
       "[5 rows x 67 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>ncl[0]</th>\n",
       "      <th>ncl[1]</th>\n",
       "      <th>ncl[2]</th>\n",
       "      <th>ncl[3]</th>\n",
       "      <th>avg_cs[0]</th>\n",
       "      <th>avg_cs[1]</th>\n",
       "      <th>avg_cs[2]</th>\n",
       "      <th>avg_cs[3]</th>\n",
       "      <th>ndof</th>\n",
       "      <th>MatchedHit_TYPE[0]</th>\n",
       "      <th>...</th>\n",
       "      <th>Mextra_DX2[1]</th>\n",
       "      <th>Mextra_DX2[2]</th>\n",
       "      <th>Mextra_DX2[3]</th>\n",
       "      <th>Mextra_DY2[0]</th>\n",
       "      <th>Mextra_DY2[1]</th>\n",
       "      <th>Mextra_DY2[2]</th>\n",
       "      <th>Mextra_DY2[3]</th>\n",
       "      <th>FOI_hits_N</th>\n",
       "      <th>PT</th>\n",
       "      <th>P</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</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",
       "      <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>0</th>\n",
       "      <td>55</td>\n",
       "      <td>12</td>\n",
       "      <td>10</td>\n",
       "      <td>15</td>\n",
       "      <td>2.127273</td>\n",
       "      <td>1.166667</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.600000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>3431.287800</td>\n",
       "      <td>7550.342300</td>\n",
       "      <td>14739.10600</td>\n",
       "      <td>1210.756600</td>\n",
       "      <td>3462.087200</td>\n",
       "      <td>7618.11100</td>\n",
       "      <td>14871.39550</td>\n",
       "      <td>6</td>\n",
       "      <td>1174.484530</td>\n",
       "      <td>6884.502272</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>43</td>\n",
       "      <td>18</td>\n",
       "      <td>12</td>\n",
       "      <td>15</td>\n",
       "      <td>2.697674</td>\n",
       "      <td>1.111111</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>1.600000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>63.777058</td>\n",
       "      <td>140.748630</td>\n",
       "      <td>266.41522</td>\n",
       "      <td>21.983400</td>\n",
       "      <td>63.552402</td>\n",
       "      <td>140.25285</td>\n",
       "      <td>265.47680</td>\n",
       "      <td>7</td>\n",
       "      <td>4344.639775</td>\n",
       "      <td>48769.526891</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>90</td>\n",
       "      <td>34</td>\n",
       "      <td>19</td>\n",
       "      <td>29</td>\n",
       "      <td>1.800000</td>\n",
       "      <td>2.058824</td>\n",
       "      <td>1.210526</td>\n",
       "      <td>1.344828</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>401.787050</td>\n",
       "      <td>903.642400</td>\n",
       "      <td>1749.45590</td>\n",
       "      <td>133.357620</td>\n",
       "      <td>383.406950</td>\n",
       "      <td>862.30440</td>\n",
       "      <td>1669.42520</td>\n",
       "      <td>4</td>\n",
       "      <td>4109.805750</td>\n",
       "      <td>20296.048412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>141</td>\n",
       "      <td>43</td>\n",
       "      <td>15</td>\n",
       "      <td>16</td>\n",
       "      <td>4.609929</td>\n",
       "      <td>2.744186</td>\n",
       "      <td>1.466667</td>\n",
       "      <td>1.125000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>211.605320</td>\n",
       "      <td>471.151430</td>\n",
       "      <td>900.27890</td>\n",
       "      <td>72.982890</td>\n",
       "      <td>211.607040</td>\n",
       "      <td>471.15527</td>\n",
       "      <td>900.28613</td>\n",
       "      <td>9</td>\n",
       "      <td>940.242439</td>\n",
       "      <td>26062.046122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>22</td>\n",
       "      <td>18</td>\n",
       "      <td>8</td>\n",
       "      <td>9</td>\n",
       "      <td>1.636364</td>\n",
       "      <td>2.444444</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.555556</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>39.059680</td>\n",
       "      <td>85.982086</td>\n",
       "      <td>163.04549</td>\n",
       "      <td>13.728162</td>\n",
       "      <td>39.018433</td>\n",
       "      <td>85.89129</td>\n",
       "      <td>162.87330</td>\n",
       "      <td>8</td>\n",
       "      <td>1386.377851</td>\n",
       "      <td>59872.343332</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 65 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    ncl[0]  ncl[1]  ncl[2]  ncl[3]  avg_cs[0]  avg_cs[1]  avg_cs[2]  \\\n",
       "id                                                                    \n",
       "0       55      12      10      15   2.127273   1.166667   1.500000   \n",
       "1       43      18      12      15   2.697674   1.111111   1.333333   \n",
       "2       90      34      19      29   1.800000   2.058824   1.210526   \n",
       "3      141      43      15      16   4.609929   2.744186   1.466667   \n",
       "4       22      18       8       9   1.636364   2.444444   1.000000   \n",
       "\n",
       "    avg_cs[3]  ndof  MatchedHit_TYPE[0]      ...       Mextra_DX2[1]  \\\n",
       "id                                           ...                       \n",
       "0    1.600000     8                   2      ...         3431.287800   \n",
       "1    1.600000     8                   2      ...           63.777058   \n",
       "2    1.344828     8                   2      ...          401.787050   \n",
       "3    1.125000     8                   2      ...          211.605320   \n",
       "4    1.555556     8                   2      ...           39.059680   \n",
       "\n",
       "    Mextra_DX2[2]  Mextra_DX2[3]  Mextra_DY2[0]  Mextra_DY2[1]  Mextra_DY2[2]  \\\n",
       "id                                                                              \n",
       "0     7550.342300    14739.10600    1210.756600    3462.087200     7618.11100   \n",
       "1      140.748630      266.41522      21.983400      63.552402      140.25285   \n",
       "2      903.642400     1749.45590     133.357620     383.406950      862.30440   \n",
       "3      471.151430      900.27890      72.982890     211.607040      471.15527   \n",
       "4       85.982086      163.04549      13.728162      39.018433       85.89129   \n",
       "\n",
       "    Mextra_DY2[3]  FOI_hits_N           PT             P  \n",
       "id                                                        \n",
       "0     14871.39550           6  1174.484530   6884.502272  \n",
       "1       265.47680           7  4344.639775  48769.526891  \n",
       "2      1669.42520           4  4109.805750  20296.048412  \n",
       "3       900.28613           9   940.242439  26062.046122  \n",
       "4       162.87330           8  1386.377851  59872.343332  \n",
       "\n",
       "[5 rows x 65 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_part, validation = train_test_split(train, test_size=0.25, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = xgboost.XGBClassifier(n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=-1, nthread=None, objective='binary:logistic',\n",
       "       random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,\n",
       "       seed=None, silent=True, subsample=1)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train_part.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values,\n",
    "          train_part.label.values,\n",
    "          sample_weight=train_part.weight.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_predictions = model.predict_proba(validation.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.548005628562668"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scoring.rejection90(validation.label.values, validation_predictions, sample_weight=validation.weight.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[18:53:50] Tree method is automatically selected to be 'approx' for faster speed. to use old behavior(exact greedy algorithm on single machine), set tree_method to 'exact'\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "       colsample_bytree=1, gamma=0, learning_rate=0.1, max_delta_step=0,\n",
       "       max_depth=3, min_child_weight=1, missing=None, n_estimators=100,\n",
       "       n_jobs=-1, nthread=None, objective='binary:logistic',\n",
       "       random_state=0, reg_alpha=0, reg_lambda=1, scale_pos_weight=1,\n",
       "       seed=None, silent=True, subsample=1)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values, train.label, sample_weight=train.weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict_proba(test.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.DataFrame(data={\"prediction\": predictions}, index=test.index).to_csv(\n",
    "    \"sample_submission.csv\", index_label=utils.ID_COLUMN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[19:26:52] Tree method is automatically selected to be 'approx' for faster speed. to use old behavior(exact greedy algorithm on single machine), set tree_method to 'exact'\n"
     ]
    }
   ],
   "source": [
    "model_track_2 = xgboost.XGBClassifier(n_estimators=150, n_jobs=-1).fit(\n",
    "    train.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values, train.label, sample_weight=train.weight)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "model_track_2.save_model(\"track_2_baseline_simple_python/track_2_model.xgb\")"
   ]
  }
 ],
 "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}