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": [
    "# The datasets are available in CoCalc in ~/share/data/I-coopetition-muon-id/\n",
    "# Test\n",
    "# ! wget --content-disposition https://codalab.coresearch.club/my/datasets/download/dd6255a1-a14b-4276-9a2b-db7f360e01c7\n",
    "# Train\n",
    "# ! wget --content-disposition https://codalab.coresearch.club/my/datasets/download/3a5e940c-2382-4716-9ff7-8fbc269b98ac"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/nkazeev/.conda/envs/py3/lib/python3.6/site-packages/numpy/lib/arraysetops.py:569: FutureWarning: elementwise comparison failed; returning scalar instead, but in the future will perform elementwise comparison\n",
      "  mask |= (ar1 == a)\n"
     ]
    }
   ],
   "source": [
    "columns = utils.SIMPLE_FEATURE_COLUMNS + [\"id\", \"label\", \"weight\", \"sWeight\", \"kinWeight\"]\n",
    "# DATA_PATH = \".\"\n",
    "DATA_PATH = \"~/share/data/I-coopetition-muon-id/\"\n",
    "train = pd.read_csv(os.path.join(DATA_PATH, \"train.csv.gz\"), index_col=\"id\", usecols=columns)\n",
    "test = pd.read_csv(os.path.join(DATA_PATH, \"test-features.csv.gz\"), index_col=\"id\", usecols=utils.SIMPLE_FEATURE_COLUMNS + [\"id\"])"
   ]
  },
  {
   "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_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>sWeight</th>\n",
       "      <th>label</th>\n",
       "      <th>kinWeight</th>\n",
       "      <th>weight</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>79</td>\n",
       "      <td>18</td>\n",
       "      <td>22</td>\n",
       "      <td>10</td>\n",
       "      <td>2.392405</td>\n",
       "      <td>2.055556</td>\n",
       "      <td>1.590909</td>\n",
       "      <td>1.200000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>5501.422000</td>\n",
       "      <td>12634.98400</td>\n",
       "      <td>25386.30500</td>\n",
       "      <td>8</td>\n",
       "      <td>1230.391860</td>\n",
       "      <td>5488.168681</td>\n",
       "      <td>1.190017</td>\n",
       "      <td>1</td>\n",
       "      <td>3.00000</td>\n",
       "      <td>3.570052</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>33</td>\n",
       "      <td>14</td>\n",
       "      <td>6</td>\n",
       "      <td>15</td>\n",
       "      <td>1.575758</td>\n",
       "      <td>2.285714</td>\n",
       "      <td>1.333333</td>\n",
       "      <td>1.400000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>671.103200</td>\n",
       "      <td>1497.54480</td>\n",
       "      <td>2901.06470</td>\n",
       "      <td>4</td>\n",
       "      <td>2360.763287</td>\n",
       "      <td>15284.170042</td>\n",
       "      <td>-0.414881</td>\n",
       "      <td>1</td>\n",
       "      <td>1.85646</td>\n",
       "      <td>-0.770209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>115</td>\n",
       "      <td>53</td>\n",
       "      <td>23</td>\n",
       "      <td>25</td>\n",
       "      <td>2.843478</td>\n",
       "      <td>2.433962</td>\n",
       "      <td>1.826087</td>\n",
       "      <td>1.280000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>917.825600</td>\n",
       "      <td>2065.64230</td>\n",
       "      <td>4027.29760</td>\n",
       "      <td>4</td>\n",
       "      <td>2685.870252</td>\n",
       "      <td>13461.339980</td>\n",
       "      <td>-0.390778</td>\n",
       "      <td>1</td>\n",
       "      <td>3.00000</td>\n",
       "      <td>-1.172334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>28</td>\n",
       "      <td>8</td>\n",
       "      <td>5</td>\n",
       "      <td>15</td>\n",
       "      <td>2.964286</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.800000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>25.208195</td>\n",
       "      <td>55.87817</td>\n",
       "      <td>106.25816</td>\n",
       "      <td>7</td>\n",
       "      <td>6015.405116</td>\n",
       "      <td>74550.106103</td>\n",
       "      <td>1.218314</td>\n",
       "      <td>1</td>\n",
       "      <td>1.85646</td>\n",
       "      <td>2.261752</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>50</td>\n",
       "      <td>8</td>\n",
       "      <td>4</td>\n",
       "      <td>13</td>\n",
       "      <td>2.400000</td>\n",
       "      <td>1.250000</td>\n",
       "      <td>1.500000</td>\n",
       "      <td>1.076923</td>\n",
       "      <td>8</td>\n",
       "      <td>2</td>\n",
       "      <td>...</td>\n",
       "      <td>408.524720</td>\n",
       "      <td>917.87640</td>\n",
       "      <td>1773.50920</td>\n",
       "      <td>5</td>\n",
       "      <td>1507.013894</td>\n",
       "      <td>19206.813287</td>\n",
       "      <td>-0.387133</td>\n",
       "      <td>1</td>\n",
       "      <td>1.85646</td>\n",
       "      <td>-0.718697</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 69 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       79      18      22      10   2.392405   2.055556   1.590909   \n",
       "1       33      14       6      15   1.575758   2.285714   1.333333   \n",
       "2      115      53      23      25   2.843478   2.433962   1.826087   \n",
       "3       28       8       5      15   2.964286   1.500000   1.800000   \n",
       "4       50       8       4      13   2.400000   1.250000   1.500000   \n",
       "\n",
       "    avg_cs[3]  ndof  MatchedHit_TYPE[0]  ...  Mextra_DY2[1]  Mextra_DY2[2]  \\\n",
       "id                                       ...                                 \n",
       "0    1.200000     8                   2  ...    5501.422000    12634.98400   \n",
       "1    1.400000     8                   2  ...     671.103200     1497.54480   \n",
       "2    1.280000     8                   2  ...     917.825600     2065.64230   \n",
       "3    1.000000     8                   2  ...      25.208195       55.87817   \n",
       "4    1.076923     8                   2  ...     408.524720      917.87640   \n",
       "\n",
       "    Mextra_DY2[3]  FOI_hits_N           PT             P   sWeight  label  \\\n",
       "id                                                                          \n",
       "0     25386.30500           8  1230.391860   5488.168681  1.190017      1   \n",
       "1      2901.06470           4  2360.763287  15284.170042 -0.414881      1   \n",
       "2      4027.29760           4  2685.870252  13461.339980 -0.390778      1   \n",
       "3       106.25816           7  6015.405116  74550.106103  1.218314      1   \n",
       "4      1773.50920           5  1507.013894  19206.813287 -0.387133      1   \n",
       "\n",
       "    kinWeight    weight  \n",
       "id                       \n",
       "0     3.00000  3.570052  \n",
       "1     1.85646 -0.770209  \n",
       "2     3.00000 -1.172334  \n",
       "3     1.85646  2.261752  \n",
       "4     1.85646 -0.718697  \n",
       "\n",
       "[5 rows x 69 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>2.697675</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   2.697675   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]  Mextra_DX2[2]  \\\n",
       "id                                       ...                                 \n",
       "0    1.600000     8                   2  ...    3431.287800    7550.342300   \n",
       "1    1.600000     8                   2  ...      63.777058     140.748630   \n",
       "2    1.344828     8                   2  ...     401.787050     903.642400   \n",
       "3    1.125000     8                   2  ...     211.605320     471.151430   \n",
       "4    1.555556     8                   2  ...      39.059680      85.982086   \n",
       "\n",
       "    Mextra_DX2[3]  Mextra_DY2[0]  Mextra_DY2[1]  Mextra_DY2[2]  Mextra_DY2[3]  \\\n",
       "id                                                                              \n",
       "0     14739.10600    1210.756600    3462.087200     7618.11100    14871.39550   \n",
       "1       266.41522      21.983400      63.552402      140.25285      265.47680   \n",
       "2      1749.45590     133.357620     383.406950      862.30440     1669.42520   \n",
       "3       900.27890      72.982890     211.607040      471.15527      900.28613   \n",
       "4       163.04549      13.728162      39.018433       85.89129      162.87330   \n",
       "\n",
       "    FOI_hits_N           PT             P  \n",
       "id                                         \n",
       "0            6  1174.484530   6884.502272  \n",
       "1            7  4344.639775  48769.526891  \n",
       "2            4  4109.805750  20296.048412  \n",
       "3            9   940.242439  26062.046122  \n",
       "4            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, random_state=2342234)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = xgboost.XGBClassifier(n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
       "              importance_type='gain', interaction_constraints='',\n",
       "              learning_rate=0.300000012, max_delta_step=0, max_depth=6,\n",
       "              min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=-1, num_parallel_tree=1,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
       "              tree_method='exact', validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 8,
     "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.kinWeight.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_predictions = model.predict_proba(validation.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.5775103906760958"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "scoring.rejection90(validation.label.values, validation_predictions, sample_weight=validation.weight.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
       "              colsample_bynode=1, colsample_bytree=1, gamma=0, gpu_id=-1,\n",
       "              importance_type='gain', interaction_constraints='',\n",
       "              learning_rate=0.300000012, max_delta_step=0, max_depth=6,\n",
       "              min_child_weight=1, missing=nan, monotone_constraints='()',\n",
       "              n_estimators=100, n_jobs=-1, num_parallel_tree=1,\n",
       "              objective='binary:logistic', random_state=0, reg_alpha=0,\n",
       "              reg_lambda=1, scale_pos_weight=1, subsample=1,\n",
       "              tree_method='exact', validate_parameters=1, verbosity=None)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.fit(train.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values, train.label, sample_weight=train.kinWeight.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict_proba(test.loc[:, utils.SIMPLE_FEATURE_COLUMNS].values)[:, 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "compression_opts = dict(method='zip',\n",
    "                        archive_name='submission.csv')  \n",
    "pd.DataFrame(data={\"prediction\": predictions}, index=test.index).to_csv(\n",
    "    \"submission.zip\", index_label=utils.ID_COLUMN, compression=compression_opts)"
   ]
  }
 ],
 "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": "3.6.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}