3_Modelling.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 3 Modeling"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.1 Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyspark.sql import SparkSession\n",
    "from pyspark.sql.functions import col, countDistinct, count, when, isnan\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from pyspark.ml import Pipeline\n",
    "import numpy as np\n",
    "from itertools import chain\n",
    "pd.set_option('max_columns', None)\n",
    "\n",
    "#temporary hide error reports\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "spark = SparkSession.builder.appName(\"modelovaniNehod\").getOrCreate()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "modelData = spark.read.parquet(\"modelData\").selectExpr(\"label\", \"features\")"
   ]
  },
  {
   "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>label</th>\n",
       "      <th>features</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>(0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>(0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>(0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>(0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                           features\n",
       "0      1  (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, ...\n",
       "1      0  (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, 1.0, ...\n",
       "2      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...\n",
       "3      0  (0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, ...\n",
       "4      0  (0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 1.0, ..."
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "modelData.limit(5).toPandas()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.2 Subsets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "trainData count is 327604, testDataCount is: 140689\n"
     ]
    }
   ],
   "source": [
    "trainData, testData = modelData.randomSplit([0.7, 0.3])\n",
    "\n",
    "print(\"trainData count is {}, testDataCount is: {}\".format(trainData.count(), testData.count()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.3 Stage"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "stages = []\n",
    "\n",
    "# DECISION TREE CLASSIFIER\n",
    "from pyspark.ml.classification import DecisionTreeClassifier\n",
    "dt = DecisionTreeClassifier(labelCol=\"label\", featuresCol=\"features\", predictionCol=\"DTprediction\", rawPredictionCol=\"DTRawPrediction\", probabilityCol=\"DTProbability\")\n",
    "stages.append(dt)\n",
    "\n",
    "# RANDOM FOREST CLASSIFIER\n",
    "from pyspark.ml.classification import RandomForestClassifier\n",
    "rf = RandomForestClassifier(labelCol=\"label\", featuresCol=\"features\", predictionCol=\"RFprediction\", rawPredictionCol=\"RFRawPrediction\", probabilityCol=\"RFProbability\", numTrees=10)\n",
    "stages.append(rf)\n",
    "\n",
    "# GBT CLASSIFIER\n",
    "from pyspark.ml.classification import GBTClassifier\n",
    "gbt = GBTClassifier(labelCol=\"label\", featuresCol=\"features\", predictionCol=\"GBTprediction\", maxIter=10)\n",
    "stages.append(gbt)\n",
    "\n",
    "# LOGISTIC REGRESSION\n",
    "from pyspark.ml.classification import LogisticRegression\n",
    "lr = LogisticRegression(featuresCol=\"features\", labelCol=\"label\", predictionCol=\"LRprediction\", probabilityCol=\"LRProbability\", rawPredictionCol=\"LRRawPrediction\", family=\"binomial\", threshold=0.5)\n",
    "stages.append(lr)\n",
    "\n",
    "# NAIVE BAYES\n",
    "from pyspark.ml.classification import NaiveBayes\n",
    "nb = NaiveBayes(smoothing=1.0, modelType=\"multinomial\", featuresCol=\"features\", labelCol=\"label\", predictionCol=\"NBprediction\", probabilityCol=\"NBProbability\", rawPredictionCol=\"NBRawPprediction\")\n",
    "stages.append(nb)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.4 Pipeline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "pipeline = Pipeline(stages = stages)\n",
    "predictionModel = pipeline.fit(trainData)\n",
    "predictions = predictionModel.transform(testData)\n",
    "predictions.createOrReplaceTempView(\"predictions\")"
   ]
  },
  {
   "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>label</th>\n",
       "      <th>features</th>\n",
       "      <th>DTRawPrediction</th>\n",
       "      <th>DTProbability</th>\n",
       "      <th>DTprediction</th>\n",
       "      <th>RFRawPrediction</th>\n",
       "      <th>RFProbability</th>\n",
       "      <th>RFprediction</th>\n",
       "      <th>rawPrediction</th>\n",
       "      <th>probability</th>\n",
       "      <th>GBTprediction</th>\n",
       "      <th>LRRawPrediction</th>\n",
       "      <th>LRProbability</th>\n",
       "      <th>LRprediction</th>\n",
       "      <th>NBRawPprediction</th>\n",
       "      <th>NBProbability</th>\n",
       "      <th>NBprediction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "      <td>[135570.0, 41139.0]</td>\n",
       "      <td>[0.7671935215523827, 0.23280647844761726]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[7.960478267123499, 2.0395217328765023]</td>\n",
       "      <td>[0.7960478267123499, 0.20395217328765022]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[0.7541448031888751, -0.7541448031888751]</td>\n",
       "      <td>[0.8188075882619448, 0.18119241173805523]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[2.0837681106289097, -2.0837681106289097]</td>\n",
       "      <td>[0.8893154859024228, 0.11068451409757715]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[-134.47888486105415, -135.64499852310377]</td>\n",
       "      <td>[0.7624418236250924, 0.2375581763749077]</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "      <td>[135570.0, 41139.0]</td>\n",
       "      <td>[0.7671935215523827, 0.23280647844761726]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[7.960478267123499, 2.0395217328765023]</td>\n",
       "      <td>[0.7960478267123499, 0.20395217328765022]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[0.7541448031888751, -0.7541448031888751]</td>\n",
       "      <td>[0.8188075882619448, 0.18119241173805523]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[2.2425641539385746, -2.2425641539385746]</td>\n",
       "      <td>[0.9040072015300331, 0.09599279846996693]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[-137.98184253354054, -139.12789914245454]</td>\n",
       "      <td>[0.7587899035455822, 0.2412100964544178]</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "      <td>[135570.0, 41139.0]</td>\n",
       "      <td>[0.7671935215523827, 0.23280647844761726]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[7.960478267123499, 2.0395217328765023]</td>\n",
       "      <td>[0.7960478267123499, 0.20395217328765022]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[0.7541448031888751, -0.7541448031888751]</td>\n",
       "      <td>[0.8188075882619448, 0.18119241173805523]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[2.2425641539385746, -2.2425641539385746]</td>\n",
       "      <td>[0.9040072015300331, 0.09599279846996693]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[-137.98184253354054, -139.12789914245454]</td>\n",
       "      <td>[0.7587899035455822, 0.2412100964544178]</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "      <td>[135570.0, 41139.0]</td>\n",
       "      <td>[0.7671935215523827, 0.23280647844761726]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[7.960478267123499, 2.0395217328765023]</td>\n",
       "      <td>[0.7960478267123499, 0.20395217328765022]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[0.7541448031888751, -0.7541448031888751]</td>\n",
       "      <td>[0.8188075882619448, 0.18119241173805523]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[2.4153486767404164, -2.4153486767404164]</td>\n",
       "      <td>[0.9179902537436024, 0.08200974625639752]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[-138.16081708346212, -139.45817592135396]</td>\n",
       "      <td>[0.7853901442182522, 0.21460985578174785]</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>(1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...</td>\n",
       "      <td>[135570.0, 41139.0]</td>\n",
       "      <td>[0.7671935215523827, 0.23280647844761726]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[7.960478267123499, 2.0395217328765023]</td>\n",
       "      <td>[0.7960478267123499, 0.20395217328765022]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[0.7541448031888751, -0.7541448031888751]</td>\n",
       "      <td>[0.8188075882619448, 0.18119241173805523]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[2.0191341710633033, -2.0191341710633033]</td>\n",
       "      <td>[0.8827914508816705, 0.11720854911832945]</td>\n",
       "      <td>0.0</td>\n",
       "      <td>[-138.56886839499757, -139.34654649453566]</td>\n",
       "      <td>[0.6851794756124814, 0.3148205243875186]</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label                                           features  \\\n",
       "0      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...   \n",
       "1      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...   \n",
       "2      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...   \n",
       "3      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...   \n",
       "4      0  (1.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 1.0, ...   \n",
       "\n",
       "       DTRawPrediction                              DTProbability  \\\n",
       "0  [135570.0, 41139.0]  [0.7671935215523827, 0.23280647844761726]   \n",
       "1  [135570.0, 41139.0]  [0.7671935215523827, 0.23280647844761726]   \n",
       "2  [135570.0, 41139.0]  [0.7671935215523827, 0.23280647844761726]   \n",
       "3  [135570.0, 41139.0]  [0.7671935215523827, 0.23280647844761726]   \n",
       "4  [135570.0, 41139.0]  [0.7671935215523827, 0.23280647844761726]   \n",
       "\n",
       "   DTprediction                          RFRawPrediction  \\\n",
       "0           0.0  [7.960478267123499, 2.0395217328765023]   \n",
       "1           0.0  [7.960478267123499, 2.0395217328765023]   \n",
       "2           0.0  [7.960478267123499, 2.0395217328765023]   \n",
       "3           0.0  [7.960478267123499, 2.0395217328765023]   \n",
       "4           0.0  [7.960478267123499, 2.0395217328765023]   \n",
       "\n",
       "                               RFProbability  RFprediction  \\\n",
       "0  [0.7960478267123499, 0.20395217328765022]           0.0   \n",
       "1  [0.7960478267123499, 0.20395217328765022]           0.0   \n",
       "2  [0.7960478267123499, 0.20395217328765022]           0.0   \n",
       "3  [0.7960478267123499, 0.20395217328765022]           0.0   \n",
       "4  [0.7960478267123499, 0.20395217328765022]           0.0   \n",
       "\n",
       "                               rawPrediction  \\\n",
       "0  [0.7541448031888751, -0.7541448031888751]   \n",
       "1  [0.7541448031888751, -0.7541448031888751]   \n",
       "2  [0.7541448031888751, -0.7541448031888751]   \n",
       "3  [0.7541448031888751, -0.7541448031888751]   \n",
       "4  [0.7541448031888751, -0.7541448031888751]   \n",
       "\n",
       "                                 probability  GBTprediction  \\\n",
       "0  [0.8188075882619448, 0.18119241173805523]            0.0   \n",
       "1  [0.8188075882619448, 0.18119241173805523]            0.0   \n",
       "2  [0.8188075882619448, 0.18119241173805523]            0.0   \n",
       "3  [0.8188075882619448, 0.18119241173805523]            0.0   \n",
       "4  [0.8188075882619448, 0.18119241173805523]            0.0   \n",
       "\n",
       "                             LRRawPrediction  \\\n",
       "0  [2.0837681106289097, -2.0837681106289097]   \n",
       "1  [2.2425641539385746, -2.2425641539385746]   \n",
       "2  [2.2425641539385746, -2.2425641539385746]   \n",
       "3  [2.4153486767404164, -2.4153486767404164]   \n",
       "4  [2.0191341710633033, -2.0191341710633033]   \n",
       "\n",
       "                               LRProbability  LRprediction  \\\n",
       "0  [0.8893154859024228, 0.11068451409757715]           0.0   \n",
       "1  [0.9040072015300331, 0.09599279846996693]           0.0   \n",
       "2  [0.9040072015300331, 0.09599279846996693]           0.0   \n",
       "3  [0.9179902537436024, 0.08200974625639752]           0.0   \n",
       "4  [0.8827914508816705, 0.11720854911832945]           0.0   \n",
       "\n",
       "                             NBRawPprediction  \\\n",
       "0  [-134.47888486105415, -135.64499852310377]   \n",
       "1  [-137.98184253354054, -139.12789914245454]   \n",
       "2  [-137.98184253354054, -139.12789914245454]   \n",
       "3  [-138.16081708346212, -139.45817592135396]   \n",
       "4  [-138.56886839499757, -139.34654649453566]   \n",
       "\n",
       "                               NBProbability  NBprediction  \n",
       "0   [0.7624418236250924, 0.2375581763749077]           0.0  \n",
       "1   [0.7587899035455822, 0.2412100964544178]           0.0  \n",
       "2   [0.7587899035455822, 0.2412100964544178]           0.0  \n",
       "3  [0.7853901442182522, 0.21460985578174785]           0.0  \n",
       "4   [0.6851794756124814, 0.3148205243875186]           0.0  "
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions.limit(5).toPandas()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.5 Print the coefficients and trees"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.1 Logistic regression coefficients"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "lrModel = lr.fit(trainData)\n",
    "intercept = lrModel.intercept"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "lrPrediction = lrModel.transform(testData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "from itertools import chain\n",
    "\n",
    "attrs = sorted((attr[\"idx\"], attr[\"name\"]) for attr in \n",
    "               (chain(*lrPrediction.schema[\"features\"].metadata[\"ml_attr\"][\"attrs\"].values())\n",
    "               )\n",
    "              )\n",
    "\n",
    "pairs = [(name, lrModel.coefficients[idx]) for idx, name in attrs]\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "sorted_pairs = sorted(pairs, key = lambda p: abs(p[1]), reverse=True)\n",
    "variables = [(\"intercept\", intercept)] + sorted_pairs\n",
    "\n",
    "results = pd.DataFrame(variables, columns = [\"predictors\", \"coefficients\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>predictors</th>\n",
       "      <th>coefficients</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s chodcem</td>\n",
       "      <td>5.344105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_řidičem nemotorového vozidla</td>\n",
       "      <td>2.437363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze  po ne...</td>\n",
       "      <td>2.423277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_výhled zakryt stojícím vo...</td>\n",
       "      <td>2.135748</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_oslnen sluncem</td>\n",
       "      <td>1.991633</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_jiné špatné</td>\n",
       "      <td>1.923700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem vegetace - ...</td>\n",
       "      <td>1.604805</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_dobré</td>\n",
       "      <td>1.586984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>DRUH_NEHODY_oh_jiný druh nehody</td>\n",
       "      <td>1.575027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem okolní zást...</td>\n",
       "      <td>1.565519</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_vyvhybal se zveri</td>\n",
       "      <td>1.410917</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem průběhu kom...</td>\n",
       "      <td>1.345455</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_podélný sklon vyšší než 8 '%</td>\n",
       "      <td>1.279128</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_nebyl ovlivnen</td>\n",
       "      <td>1.156279</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_příčná stružka, hrbol, vyst...</td>\n",
       "      <td>1.088260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_ovlinen jednanim os...</td>\n",
       "      <td>1.055701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_přechodná uzavírka komunika...</td>\n",
       "      <td>0.975983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_verejna doprava</td>\n",
       "      <td>0.966913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze, stání...</td>\n",
       "      <td>0.948091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>DRUH_NEHODY_oh_havárie</td>\n",
       "      <td>0.914099</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_jiný (neuvedený) stav nebo ...</td>\n",
       "      <td>0.845669</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_jiný sta...</td>\n",
       "      <td>0.844430</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_dobrý, bez závad</td>\n",
       "      <td>0.804896</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s domácím zvířetem</td>\n",
       "      <td>0.780548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>HLAVNÍ_PŘÍČINY_NEHODY_oh_nedání přednosti v jízdě</td>\n",
       "      <td>0.771145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_z...</td>\n",
       "      <td>0.744225</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_p...</td>\n",
       "      <td>0.726046</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_zvlněný povrch v podélném s...</td>\n",
       "      <td>0.669031</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 2. třídy</td>\n",
       "      <td>0.650610</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch s...</td>\n",
       "      <td>0.646419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_náhlá zm...</td>\n",
       "      <td>0.570239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s jedoucím nekolejovým v...</td>\n",
       "      <td>0.569471</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_žádná místn...</td>\n",
       "      <td>0.561269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze po spr...</td>\n",
       "      <td>0.522446</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>DRUH_POVRCHU_VOZOVKY_oh_jiný nezpevněný povrch</td>\n",
       "      <td>0.510313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_souvislé výtluky</td>\n",
       "      <td>0.507195</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_místní úprava</td>\n",
       "      <td>0.507185</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 3. třídy</td>\n",
       "      <td>0.475030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 1. třídy</td>\n",
       "      <td>0.472567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_čtyřpruhová s dělící čarou</td>\n",
       "      <td>0.472196</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_v...</td>\n",
       "      <td>0.465540</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_přímý úsek po projetí zatáčkou</td>\n",
       "      <td>0.457988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_mlha</td>\n",
       "      <td>0.426121</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_z...</td>\n",
       "      <td>0.422757</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>0.405002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_křižovatka průsečná - čtyřra...</td>\n",
       "      <td>0.397429</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>76</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_v...</td>\n",
       "      <td>0.371209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>77</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_soukrome</td>\n",
       "      <td>0.370285</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>78</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_přecházení  ...</td>\n",
       "      <td>0.364105</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>0.319013</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>83</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_nesouvislé výtluky</td>\n",
       "      <td>0.311874</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>85</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...</td>\n",
       "      <td>0.305092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>86</th>\n",
       "      <td>DRUH_POVRCHU_VOZOVKY_oh_beton</td>\n",
       "      <td>0.301881</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_dvoupruhová</td>\n",
       "      <td>0.290634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>88</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_ž...</td>\n",
       "      <td>0.274561</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93</th>\n",
       "      <td>HLAVNÍ_PŘÍČINY_NEHODY_oh_nepřiměřená rychlost ...</td>\n",
       "      <td>0.255916</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>94</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_křižovatka styková - tříramenná</td>\n",
       "      <td>0.252213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>98</th>\n",
       "      <td>MÍSTO_DOPRAVNÍ_NEHODY_oh_mimo křižovatku</td>\n",
       "      <td>0.234145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>99</th>\n",
       "      <td>DRUH_POVRCHU_VOZOVKY_oh_živice</td>\n",
       "      <td>0.233035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100</th>\n",
       "      <td>HLAVNÍ_PŘÍČINY_NEHODY_oh_nesprávné předjíždění</td>\n",
       "      <td>0.232755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_sněžení</td>\n",
       "      <td>0.226747</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104</th>\n",
       "      <td>VIDITELNOST_oh_v noci bez veřejného osvětlení,...</td>\n",
       "      <td>0.223003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>105</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch s...</td>\n",
       "      <td>0.213061</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>106</th>\n",
       "      <td>POČET_ZÚČASTNĚNÝCH_VOZIDEL_oh_2.0</td>\n",
       "      <td>0.211434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>107</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_tvoří s...</td>\n",
       "      <td>0.202505</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>110</th>\n",
       "      <td>DRUH_PEVNÉ_PŘEKÁŽKY_oh_ne</td>\n",
       "      <td>0.190371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>112</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_na počá...</td>\n",
       "      <td>0.178411</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>113</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_déšť</td>\n",
       "      <td>0.174164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>114</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s pevnou překážkou</td>\n",
       "      <td>0.173623</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>115</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_kruhový objezd</td>\n",
       "      <td>0.173357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>117</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_uzel  tj. křižovatk...</td>\n",
       "      <td>0.171317</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>118</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ne</td>\n",
       "      <td>0.166777</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>119</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>0.166651</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>123</th>\n",
       "      <td>MÍSTO_DOPRAVNÍ_NEHODY_oh_na kžižovatce jedná-l...</td>\n",
       "      <td>0.154760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>124</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_neztížené</td>\n",
       "      <td>0.149047</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>VIDITELNOST_oh_ve dne viditelnost nezhoršená v...</td>\n",
       "      <td>0.147221</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>127</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...</td>\n",
       "      <td>0.139413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>128</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_nezjišťováno</td>\n",
       "      <td>0.137300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>130</th>\n",
       "      <td>OBDOBI_oh_leto</td>\n",
       "      <td>0.123044</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>132</th>\n",
       "      <td>DRUH_POVRCHU_VOZOVKY_oh_štěrk</td>\n",
       "      <td>0.106814</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>135</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch m...</td>\n",
       "      <td>0.100204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>136</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...</td>\n",
       "      <td>0.098792</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>137</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_zatáčka</td>\n",
       "      <td>0.094182</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>139</th>\n",
       "      <td>HLAVNÍ_PŘÍČINY_NEHODY_oh_nezaviněná řidičem</td>\n",
       "      <td>0.080659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>141</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_pod vlivem...</td>\n",
       "      <td>0.076999</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>143</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_vícepruhová</td>\n",
       "      <td>0.065040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>146</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_čtyřpruhová s dělícím pásem</td>\n",
       "      <td>0.045693</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>147</th>\n",
       "      <td>VIDITELNOST_oh_ve dne zhoršená viditelnost (sv...</td>\n",
       "      <td>0.041469</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>148</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_m...</td>\n",
       "      <td>0.040101</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>149</th>\n",
       "      <td>SMĚROVÉ_POMĚRY_oh_přímý úsek</td>\n",
       "      <td>0.038967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>150</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...</td>\n",
       "      <td>0.038557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>151</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>0.035936</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>152</th>\n",
       "      <td>VIDITELNOST_oh_v noci bez veřejného osvětlení,...</td>\n",
       "      <td>0.034873</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>153</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_třípruhová</td>\n",
       "      <td>0.034592</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>160</th>\n",
       "      <td>DEN_V_TYDNU_oh_predPoVikend</td>\n",
       "      <td>0.014404</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>161</th>\n",
       "      <td>VIDITELNOST_oh_ve dne zhoršená viditelnost vli...</td>\n",
       "      <td>0.013270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>164</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_firemni</td>\n",
       "      <td>0.002980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>163</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_č...</td>\n",
       "      <td>-0.003436</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>162</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_ž...</td>\n",
       "      <td>-0.008198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>159</th>\n",
       "      <td>MÍSTO_DOPRAVNÍ_NEHODY_oh_na křižovatce uvnitř ...</td>\n",
       "      <td>-0.016295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>158</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...</td>\n",
       "      <td>-0.019983</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>157</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_řidičem motorového vozidla</td>\n",
       "      <td>-0.022588</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>156</th>\n",
       "      <td>DEN_V_TYDNU_oh_tyden</td>\n",
       "      <td>-0.023102</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>155</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_jiná situace</td>\n",
       "      <td>-0.024622</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>154</th>\n",
       "      <td>ČAS_oh_den</td>\n",
       "      <td>-0.028834</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>145</th>\n",
       "      <td>ČAS_oh_vecer</td>\n",
       "      <td>-0.051836</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>144</th>\n",
       "      <td>DRUH_POVRCHU_VOZOVKY_oh_dlažba</td>\n",
       "      <td>-0.053350</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_nesprávně umístěná, znečišt...</td>\n",
       "      <td>-0.076646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>140</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>-0.080184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>138</th>\n",
       "      <td>SMYK_oh_ne</td>\n",
       "      <td>-0.092133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>134</th>\n",
       "      <td>OBDOBI_oh_jaro</td>\n",
       "      <td>-0.100435</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>133</th>\n",
       "      <td>VIDITELNOST_oh_v noci s veřejným osvětlením, v...</td>\n",
       "      <td>-0.104700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>131</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace místní</td>\n",
       "      <td>-0.112662</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>129</th>\n",
       "      <td>CELKOVÁ_HMOTNÁ_ŠKODA_oh_1.0</td>\n",
       "      <td>-0.130270</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>126</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_dálnice</td>\n",
       "      <td>-0.139593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>122</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_mimo CR</td>\n",
       "      <td>-0.157615</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>121</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_přecházení p...</td>\n",
       "      <td>-0.159343</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>120</th>\n",
       "      <td>SVATEK_oh_Ne</td>\n",
       "      <td>-0.162218</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>116</th>\n",
       "      <td>MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost vy...</td>\n",
       "      <td>-0.172897</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>111</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace sledovan...</td>\n",
       "      <td>-0.183005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>109</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_jine</td>\n",
       "      <td>-0.191573</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>108</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s tramvají</td>\n",
       "      <td>-0.199956</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103</th>\n",
       "      <td>STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_souvislá...</td>\n",
       "      <td>-0.225348</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_nakladni doprava</td>\n",
       "      <td>-0.229743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>97</th>\n",
       "      <td>ROK_VÝROBY_VOZIDLA_oh_1.0</td>\n",
       "      <td>-0.237608</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_přechodná uzavírka jednoho ...</td>\n",
       "      <td>-0.243269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>95</th>\n",
       "      <td>MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost vy...</td>\n",
       "      <td>-0.244901</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>92</th>\n",
       "      <td>HLAVNÍ_PŘÍČINY_NEHODY_oh_nesprávný způsob jízdy</td>\n",
       "      <td>-0.258137</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>91</th>\n",
       "      <td>MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost ne...</td>\n",
       "      <td>-0.264376</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90</th>\n",
       "      <td>OBDOBI_oh_zima</td>\n",
       "      <td>-0.264495</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>89</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_odcizene</td>\n",
       "      <td>-0.269535</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>84</th>\n",
       "      <td>ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_světelným sign...</td>\n",
       "      <td>-0.310685</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>81</th>\n",
       "      <td>SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_p...</td>\n",
       "      <td>-0.340835</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>80</th>\n",
       "      <td>DĚLENÍ_KOMUNIKACE_oh_žádná z uvedených</td>\n",
       "      <td>-0.355381</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>79</th>\n",
       "      <td>POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_nárazov...</td>\n",
       "      <td>-0.355590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75</th>\n",
       "      <td>ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_žádný způsob ř...</td>\n",
       "      <td>-0.372637</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>74</th>\n",
       "      <td>SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...</td>\n",
       "      <td>-0.383966</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>LOKALITA_NEHODY_oh_v obci</td>\n",
       "      <td>-0.425222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>KATEGORIE_ŘIDIČE_oh_s ridicskym opravnenim</td>\n",
       "      <td>-0.472258</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>ROK_VÝROBY_VOZIDLA_oh_2.0</td>\n",
       "      <td>-0.513522</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...</td>\n",
       "      <td>-0.544028</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_zezadu</td>\n",
       "      <td>-0.554005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>ŠKODA_NA_VOZIDLE_oh_1.0</td>\n",
       "      <td>-0.634579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_chodcem</td>\n",
       "      <td>-0.649837</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_nepřichází v ...</td>\n",
       "      <td>-0.657311</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace účelová ...</td>\n",
       "      <td>-0.691346</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>STAV_ŘIDIČE_oh_jiny</td>\n",
       "      <td>-0.709431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>STAV_KOMUNIKACE_oh_neoznačená nebo nedostatečn...</td>\n",
       "      <td>-0.709853</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_z boku</td>\n",
       "      <td>-0.900108</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_jiné zavinění</td>\n",
       "      <td>-0.987892</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s vozidlem zaparkovaným,...</td>\n",
       "      <td>-1.005328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_technickou závadou vozidla</td>\n",
       "      <td>-1.062660</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>CHARAKTERISTIKA_VOZIDLA__oh_statni</td>\n",
       "      <td>-1.142002</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_boční</td>\n",
       "      <td>-1.192909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>STAV_ŘIDIČE_oh_pod vlivem navykovych latek</td>\n",
       "      <td>-1.269961</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>intercept</td>\n",
       "      <td>-1.428533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_nevime</td>\n",
       "      <td>-1.431009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>STAV_ŘIDIČE_oh_dobry</td>\n",
       "      <td>-1.452160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_lesní zvěří, domácím zvířec...</td>\n",
       "      <td>-1.611431</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>DRUH_NEHODY_oh_srážka s lesní zvěří</td>\n",
       "      <td>-1.899671</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>DRUH_VOZIDLA_oh_ostatni</td>\n",
       "      <td>-1.909391</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>DRUH_VOZIDLA_oh_dopravni</td>\n",
       "      <td>-2.119316</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ZAVINĚNÍ_NEHODY_oh_závadou komunikace</td>\n",
       "      <td>-2.611634</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DRUH_VOZIDLA_oh_osobni</td>\n",
       "      <td>-2.784040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>DRUH_VOZIDLA_oh_nakladni</td>\n",
       "      <td>-2.874833</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                            predictors  coefficients\n",
       "1                      DRUH_NEHODY_oh_srážka s chodcem      5.344105\n",
       "5      ZAVINĚNÍ_NEHODY_oh_řidičem nemotorového vozidla      2.437363\n",
       "6    SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze  po ne...      2.423277\n",
       "7    ROZHLEDOVÉ_POMĚRY_oh_výhled zakryt stojícím vo...      2.135748\n",
       "9            VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_oslnen sluncem      1.991633\n",
       "10                   ROZHLEDOVÉ_POMĚRY_oh_jiné špatné       1.923700\n",
       "14   ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem vegetace - ...      1.604805\n",
       "15                         ROZHLEDOVÉ_POMĚRY_oh_dobré       1.586984\n",
       "16                    DRUH_NEHODY_oh_jiný druh nehody       1.575027\n",
       "17   ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem okolní zást...      1.565519\n",
       "20        VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_vyvhybal se zveri      1.410917\n",
       "21   ROZHLEDOVÉ_POMĚRY_oh_špatné vlivem průběhu kom...      1.345455\n",
       "22     STAV_KOMUNIKACE_oh_podélný sklon vyšší než 8 '%      1.279128\n",
       "25           VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_nebyl ovlivnen      1.156279\n",
       "27   STAV_KOMUNIKACE_oh_příčná stružka, hrbol, vyst...      1.088260\n",
       "29   VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_ovlinen jednanim os...      1.055701\n",
       "32   STAV_KOMUNIKACE_oh_přechodná uzavírka komunika...      0.975983\n",
       "33         CHARAKTERISTIKA_VOZIDLA__oh_verejna doprava      0.966913\n",
       "34   SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze, stání...      0.948091\n",
       "35                             DRUH_NEHODY_oh_havárie       0.914099\n",
       "37   STAV_KOMUNIKACE_oh_jiný (neuvedený) stav nebo ...      0.845669\n",
       "38   STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_jiný sta...      0.844430\n",
       "39                 STAV_KOMUNIKACE_oh_dobrý, bez závad      0.804896\n",
       "40            DRUH_NEHODY_oh_srážka s domácím zvířetem      0.780548\n",
       "41   HLAVNÍ_PŘÍČINY_NEHODY_oh_nedání přednosti v jízdě      0.771145\n",
       "42   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_z...      0.744225\n",
       "43   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_p...      0.726046\n",
       "47   STAV_KOMUNIKACE_oh_zvlněný povrch v podélném s...      0.669031\n",
       "49        DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 2. třídy       0.650610\n",
       "51   STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch s...      0.646419\n",
       "53   STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_náhlá zm...      0.570239\n",
       "54   DRUH_NEHODY_oh_srážka s jedoucím nekolejovým v...      0.569471\n",
       "55   MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_žádná místn...      0.561269\n",
       "58   SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_chůze po spr...      0.522446\n",
       "60      DRUH_POVRCHU_VOZOVKY_oh_jiný nezpevněný povrch      0.510313\n",
       "61                 STAV_KOMUNIKACE_oh_souvislé výtluky      0.507195\n",
       "62      ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_místní úprava       0.507185\n",
       "63        DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 3. třídy       0.475030\n",
       "64        DRUH_POZEMNÍ_KOMUNIKACE_oh_silnice 1. třídy       0.472567\n",
       "66     DĚLENÍ_KOMUNIKACE_oh_čtyřpruhová s dělící čarou      0.472196\n",
       "67   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_v...      0.465540\n",
       "68   SMĚROVÉ_POMĚRY_oh_přímý úsek po projetí zatáčkou       0.457988\n",
       "69         POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_mlha      0.426121\n",
       "71   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_z...      0.422757\n",
       "72   ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...      0.405002\n",
       "73   SMĚROVÉ_POMĚRY_oh_křižovatka průsečná - čtyřra...      0.397429\n",
       "76   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_v...      0.371209\n",
       "77                CHARAKTERISTIKA_VOZIDLA__oh_soukrome      0.370285\n",
       "78   SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_přecházení  ...      0.364105\n",
       "82   ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...      0.319013\n",
       "83               STAV_KOMUNIKACE_oh_nesouvislé výtluky      0.311874\n",
       "85   STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...      0.305092\n",
       "86                       DRUH_POVRCHU_VOZOVKY_oh_beton      0.301881\n",
       "87                    DĚLENÍ_KOMUNIKACE_oh_dvoupruhová      0.290634\n",
       "88   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_ž...      0.274561\n",
       "93   HLAVNÍ_PŘÍČINY_NEHODY_oh_nepřiměřená rychlost ...      0.255916\n",
       "94   SMĚROVÉ_POMĚRY_oh_křižovatka styková - tříramenná      0.252213\n",
       "98           MÍSTO_DOPRAVNÍ_NEHODY_oh_mimo křižovatku       0.234145\n",
       "99                      DRUH_POVRCHU_VOZOVKY_oh_živice      0.233035\n",
       "100     HLAVNÍ_PŘÍČINY_NEHODY_oh_nesprávné předjíždění      0.232755\n",
       "102     POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_sněžení      0.226747\n",
       "104  VIDITELNOST_oh_v noci bez veřejného osvětlení,...      0.223003\n",
       "105  STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch s...      0.213061\n",
       "106                  POČET_ZÚČASTNĚNÝCH_VOZIDEL_oh_2.0      0.211434\n",
       "107  POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_tvoří s...      0.202505\n",
       "110                          DRUH_PEVNÉ_PŘEKÁŽKY_oh_ne      0.190371\n",
       "112  POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_na počá...      0.178411\n",
       "113        POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_déšť      0.174164\n",
       "114           DRUH_NEHODY_oh_srážka s pevnou překážkou      0.173623\n",
       "115                   SMĚROVÉ_POMĚRY_oh_kruhový objezd      0.173357\n",
       "117  DRUH_POZEMNÍ_KOMUNIKACE_oh_uzel  tj. křižovatk...      0.171317\n",
       "118            ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ne       0.166777\n",
       "119  ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...      0.166651\n",
       "123  MÍSTO_DOPRAVNÍ_NEHODY_oh_na kžižovatce jedná-l...      0.154760\n",
       "124   POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_neztížené      0.149047\n",
       "125  VIDITELNOST_oh_ve dne viditelnost nezhoršená v...      0.147221\n",
       "127  SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...      0.139413\n",
       "128  ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_nezjišťováno       0.137300\n",
       "130                                     OBDOBI_oh_leto      0.123044\n",
       "132                      DRUH_POVRCHU_VOZOVKY_oh_štěrk      0.106814\n",
       "135  STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_povrch m...      0.100204\n",
       "136  SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...      0.098792\n",
       "137                          SMĚROVÉ_POMĚRY_oh_zatáčka      0.094182\n",
       "139        HLAVNÍ_PŘÍČINY_NEHODY_oh_nezaviněná řidičem      0.080659\n",
       "141  ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_pod vlivem...      0.076999\n",
       "143                   DĚLENÍ_KOMUNIKACE_oh_vícepruhová      0.065040\n",
       "146   DĚLENÍ_KOMUNIKACE_oh_čtyřpruhová s dělícím pásem      0.045693\n",
       "147  VIDITELNOST_oh_ve dne zhoršená viditelnost (sv...      0.041469\n",
       "148  SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_m...      0.040101\n",
       "149                       SMĚROVÉ_POMĚRY_oh_přímý úsek      0.038967\n",
       "150  STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...      0.038557\n",
       "151  ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...      0.035936\n",
       "152  VIDITELNOST_oh_v noci bez veřejného osvětlení,...      0.034873\n",
       "153                    DĚLENÍ_KOMUNIKACE_oh_třípruhová      0.034592\n",
       "160                        DEN_V_TYDNU_oh_predPoVikend      0.014404\n",
       "161  VIDITELNOST_oh_ve dne zhoršená viditelnost vli...      0.013270\n",
       "164                CHARAKTERISTIKA_VOZIDLA__oh_firemni      0.002980\n",
       "163  SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_č...     -0.003436\n",
       "162  SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_ž...     -0.008198\n",
       "159  MÍSTO_DOPRAVNÍ_NEHODY_oh_na křižovatce uvnitř ...     -0.016295\n",
       "158  STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_na vozov...     -0.019983\n",
       "157      ZAVINĚNÍ_NEHODY_oh_řidičem motorového vozidla     -0.022588\n",
       "156                               DEN_V_TYDNU_oh_tyden     -0.023102\n",
       "155    SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_jiná situace      -0.024622\n",
       "154                                         ČAS_oh_den     -0.028834\n",
       "145                                       ČAS_oh_vecer     -0.051836\n",
       "144                     DRUH_POVRCHU_VOZOVKY_oh_dlažba     -0.053350\n",
       "142  STAV_KOMUNIKACE_oh_nesprávně umístěná, znečišt...     -0.076646\n",
       "140  ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...     -0.080184\n",
       "138                                         SMYK_oh_ne     -0.092133\n",
       "134                                     OBDOBI_oh_jaro     -0.100435\n",
       "133  VIDITELNOST_oh_v noci s veřejným osvětlením, v...     -0.104700\n",
       "131      DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace místní      -0.112662\n",
       "129                        CELKOVÁ_HMOTNÁ_ŠKODA_oh_1.0     -0.130270\n",
       "126                DRUH_POZEMNÍ_KOMUNIKACE_oh_dálnice      -0.139593\n",
       "122                CHARAKTERISTIKA_VOZIDLA__oh_mimo CR     -0.157615\n",
       "121  SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_přecházení p...     -0.159343\n",
       "120                                       SVATEK_oh_Ne     -0.162218\n",
       "116  MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost vy...     -0.172897\n",
       "111  DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace sledovan...     -0.183005\n",
       "109                   CHARAKTERISTIKA_VOZIDLA__oh_jine     -0.191573\n",
       "108                   DRUH_NEHODY_oh_srážka s tramvají     -0.199956\n",
       "103  STAV_POVRCHU_VOZOVKY_V_DOBĚ_NEHODY_oh_souvislá...     -0.225348\n",
       "101       CHARAKTERISTIKA_VOZIDLA__oh_nakladni doprava     -0.229743\n",
       "97                           ROK_VÝROBY_VOZIDLA_oh_1.0     -0.237608\n",
       "96   STAV_KOMUNIKACE_oh_přechodná uzavírka jednoho ...     -0.243269\n",
       "95   MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost vy...     -0.244901\n",
       "92     HLAVNÍ_PŘÍČINY_NEHODY_oh_nesprávný způsob jízdy     -0.258137\n",
       "91   MÍSTNÍ_ÚPRAVA_PŘEDNOSTI_V_JÍZDĚ_oh_přednost ne...     -0.264376\n",
       "90                                      OBDOBI_oh_zima     -0.264495\n",
       "89                CHARAKTERISTIKA_VOZIDLA__oh_odcizene     -0.269535\n",
       "84   ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_světelným sign...     -0.310685\n",
       "81   SPECIFICKÁ_MÍSTA_A_OBJEKTY_V_MÍSTĚ_NEHODY_oh_p...     -0.340835\n",
       "80              DĚLENÍ_KOMUNIKACE_oh_žádná z uvedených     -0.355381\n",
       "79   POVĚTRNOSTNÍ_PODMÍNKY_V_DOBĚ_NEHODY_oh_nárazov...     -0.355590\n",
       "75   ŘÍZENÍ_PROVOZU_V_DOBĚ_NEHODY_oh_žádný způsob ř...     -0.372637\n",
       "74   SITUOVÁNÍ_NEHODY_NA_KOMUNIKACI_oh_vstup chodce...     -0.383966\n",
       "70                           LOKALITA_NEHODY_oh_v obci     -0.425222\n",
       "65          KATEGORIE_ŘIDIČE_oh_s ridicskym opravnenim     -0.472258\n",
       "59                           ROK_VÝROBY_VOZIDLA_oh_2.0     -0.513522\n",
       "57   ALKOHOL_U_VINÍKA_NEHODY_PŘÍTOMEN_oh_ano obsah ...     -0.544028\n",
       "56             DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_zezadu     -0.554005\n",
       "52                             ŠKODA_NA_VOZIDLE_oh_1.0     -0.634579\n",
       "50                         ZAVINĚNÍ_NEHODY_oh_chodcem      -0.649837\n",
       "48   DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_nepřichází v ...     -0.657311\n",
       "46   DRUH_POZEMNÍ_KOMUNIKACE_oh_komunikace účelová ...     -0.691346\n",
       "45                                 STAV_ŘIDIČE_oh_jiny     -0.709431\n",
       "44   STAV_KOMUNIKACE_oh_neoznačená nebo nedostatečn...     -0.709853\n",
       "36             DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_z boku     -0.900108\n",
       "31                   ZAVINĚNÍ_NEHODY_oh_jiné zavinění      -0.987892\n",
       "30   DRUH_NEHODY_oh_srážka s vozidlem zaparkovaným,...     -1.005328\n",
       "28      ZAVINĚNÍ_NEHODY_oh_technickou závadou vozidla      -1.062660\n",
       "26                  CHARAKTERISTIKA_VOZIDLA__oh_statni     -1.142002\n",
       "24              DRUH_SRÁŽKY_JEDOUCÍCH_VOZIDEL_oh_boční     -1.192909\n",
       "23          STAV_ŘIDIČE_oh_pod vlivem navykovych latek     -1.269961\n",
       "0                                            intercept     -1.428533\n",
       "19                   VNĚJŠÍ_OVLIVNĚNÍ_ŘIDIČE_oh_nevime     -1.431009\n",
       "18                                STAV_ŘIDIČE_oh_dobry     -1.452160\n",
       "13   ZAVINĚNÍ_NEHODY_oh_lesní zvěří, domácím zvířec...     -1.611431\n",
       "12                 DRUH_NEHODY_oh_srážka s lesní zvěří     -1.899671\n",
       "11                             DRUH_VOZIDLA_oh_ostatni     -1.909391\n",
       "8                             DRUH_VOZIDLA_oh_dopravni     -2.119316\n",
       "4               ZAVINĚNÍ_NEHODY_oh_závadou komunikace      -2.611634\n",
       "3                               DRUH_VOZIDLA_oh_osobni     -2.784040\n",
       "2                             DRUH_VOZIDLA_oh_nakladni     -2.874833"
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.set_option('display.max_rows', None)\n",
    "results.sort_values('coefficients', ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# export Logit coefficients to csv (for visualization)\n",
    "results.to_csv('lrCoefficients.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.2 Decision tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DecisionTreeClassificationModel (uid=DecisionTreeClassifier_28e3a49ca62f) of depth 5 with 35 nodes\n",
      "  If (feature 104 in {1.0})\n",
      "   If (feature 130 in {1.0})\n",
      "    If (feature 8 in {1.0})\n",
      "     If (feature 161 in {1.0})\n",
      "      If (feature 35 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 35 not in {1.0})\n",
      "       Predict: 1.0\n",
      "     Else (feature 161 not in {1.0})\n",
      "      If (feature 18 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 18 not in {1.0})\n",
      "       Predict: 1.0\n",
      "    Else (feature 8 not in {1.0})\n",
      "     If (feature 147 in {1.0})\n",
      "      If (feature 10 in {0.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 10 not in {0.0})\n",
      "       Predict: 1.0\n",
      "     Else (feature 147 not in {1.0})\n",
      "      If (feature 117 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 117 not in {1.0})\n",
      "       Predict: 1.0\n",
      "   Else (feature 130 not in {1.0})\n",
      "    If (feature 141 in {0.0})\n",
      "     If (feature 56 in {1.0})\n",
      "      Predict: 0.0\n",
      "     Else (feature 56 not in {1.0})\n",
      "      Predict: 1.0\n",
      "    Else (feature 141 not in {0.0})\n",
      "     Predict: 1.0\n",
      "  Else (feature 104 not in {1.0})\n",
      "   If (feature 121 in {1.0})\n",
      "    If (feature 115 in {1.0})\n",
      "     If (feature 156 in {1.0})\n",
      "      If (feature 141 in {0.0})\n",
      "       Predict: 1.0\n",
      "      Else (feature 141 not in {0.0})\n",
      "       Predict: 0.0\n",
      "     Else (feature 156 not in {1.0})\n",
      "      If (feature 100 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 100 not in {1.0})\n",
      "       Predict: 1.0\n",
      "    Else (feature 115 not in {1.0})\n",
      "     Predict: 0.0\n",
      "   Else (feature 121 not in {1.0})\n",
      "    If (feature 115 in {1.0})\n",
      "     Predict: 1.0\n",
      "    Else (feature 115 not in {1.0})\n",
      "     Predict: 0.0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "treeModel = dt.fit(testData)\n",
    "print(treeModel.toDebugString)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.3 Random forests"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RandomForestClassificationModel (uid=RandomForestClassifier_1defcfb00d6c) with 10 trees\n",
      "  Tree 0 (weight 1.0):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 149 in {1.0})\n",
      "      If (feature 57 in {0.0})\n",
      "       If (feature 161 in {0.0})\n",
      "        If (feature 170 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 170 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 161 not in {0.0})\n",
      "        Predict: 0.0\n",
      "      Else (feature 57 not in {0.0})\n",
      "       If (feature 23 in {0.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 23 not in {0.0})\n",
      "        If (feature 101 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 101 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 149 not in {1.0})\n",
      "      Predict: 0.0\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     Predict: 1.0\n",
      "  Tree 1 (weight 1.0):\n",
      "    If (feature 42 in {1.0})\n",
      "     If (feature 7 in {1.0})\n",
      "      If (feature 2 <= 0.7233277630270138)\n",
      "       Predict: 0.0\n",
      "      Else (feature 2 > 0.7233277630270138)\n",
      "       Predict: 1.0\n",
      "     Else (feature 7 not in {1.0})\n",
      "      If (feature 140 in {1.0})\n",
      "       If (feature 6 <= 0.1070547951588865)\n",
      "        If (feature 149 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 149 not in {1.0})\n",
      "         Predict: 0.0\n",
      "       Else (feature 6 > 0.1070547951588865)\n",
      "        Predict: 0.0\n",
      "      Else (feature 140 not in {1.0})\n",
      "       If (feature 172 in {0.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 172 not in {0.0})\n",
      "        If (feature 160 in {0.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 160 not in {0.0})\n",
      "         Predict: 0.0\n",
      "    Else (feature 42 not in {1.0})\n",
      "     If (feature 129 in {1.0})\n",
      "      If (feature 43 in {1.0})\n",
      "       Predict: 1.0\n",
      "      Else (feature 43 not in {1.0})\n",
      "       If (feature 58 in {1.0})\n",
      "        If (feature 24 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 24 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 58 not in {1.0})\n",
      "        Predict: 0.0\n",
      "     Else (feature 129 not in {1.0})\n",
      "      If (feature 149 in {1.0})\n",
      "       Predict: 1.0\n",
      "      Else (feature 149 not in {1.0})\n",
      "       If (feature 131 in {1.0})\n",
      "        If (feature 20 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 20 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 131 not in {1.0})\n",
      "        Predict: 0.0\n",
      "  Tree 2 (weight 1.0):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 24 in {1.0})\n",
      "      If (feature 59 in {1.0})\n",
      "       Predict: 1.0\n",
      "      Else (feature 59 not in {1.0})\n",
      "       If (feature 49 in {1.0})\n",
      "        If (feature 140 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 140 not in {1.0})\n",
      "         Predict: 0.0\n",
      "       Else (feature 49 not in {1.0})\n",
      "        If (feature 129 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 129 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 24 not in {1.0})\n",
      "      If (feature 147 in {0.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 147 not in {0.0})\n",
      "       If (feature 28 in {1.0})\n",
      "        If (feature 0 <= 6.777901014488855)\n",
      "         Predict: 0.0\n",
      "        Else (feature 0 > 6.777901014488855)\n",
      "         Predict: 1.0\n",
      "       Else (feature 28 not in {1.0})\n",
      "        Predict: 0.0\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     Predict: 1.0\n",
      "  Tree 3 (weight 1.0):\n",
      "    If (feature 58 in {1.0})\n",
      "     If (feature 94 in {0.0})\n",
      "      If (feature 123 in {1.0})\n",
      "       If (feature 25 in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "       Else (feature 25 not in {1.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 123 not in {1.0})\n",
      "       If (feature 104 in {1.0})\n",
      "        Predict: 1.0\n",
      "       Else (feature 104 not in {1.0})\n",
      "        If (feature 24 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 24 not in {0.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 94 not in {0.0})\n",
      "      If (feature 124 in {0.0})\n",
      "       If (feature 117 in {1.0})\n",
      "        Predict: 1.0\n",
      "       Else (feature 117 not in {1.0})\n",
      "        Predict: 0.0\n",
      "      Else (feature 124 not in {0.0})\n",
      "       If (feature 49 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 49 not in {1.0})\n",
      "        Predict: 1.0\n",
      "    Else (feature 58 not in {1.0})\n",
      "     If (feature 149 in {1.0})\n",
      "      If (feature 38 in {1.0})\n",
      "       If (feature 43 in {0.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 43 not in {0.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 38 not in {1.0})\n",
      "       If (feature 162 in {1.0})\n",
      "        If (feature 27 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 27 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 162 not in {1.0})\n",
      "        If (feature 20 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 20 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 149 not in {1.0})\n",
      "      If (feature 122 in {1.0})\n",
      "       If (feature 2 <= 0.7233277630270138)\n",
      "        Predict: 0.0\n",
      "       Else (feature 2 > 0.7233277630270138)\n",
      "        Predict: 1.0\n",
      "      Else (feature 122 not in {1.0})\n",
      "       If (feature 142 in {1.0})\n",
      "        If (feature 0 <= 6.777901014488855)\n",
      "         Predict: 0.0\n",
      "        Else (feature 0 > 6.777901014488855)\n",
      "         Predict: 1.0\n",
      "       Else (feature 142 not in {1.0})\n",
      "        Predict: 0.0\n",
      "  Tree 4 (weight 1.0):\n",
      "    If (feature 24 in {1.0})\n",
      "     If (feature 44 in {1.0})\n",
      "      If (feature 155 in {1.0})\n",
      "       If (feature 17 in {0.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 17 not in {0.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 155 not in {1.0})\n",
      "       If (feature 129 in {1.0})\n",
      "        If (feature 121 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 121 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 129 not in {1.0})\n",
      "        If (feature 132 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 132 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 44 not in {1.0})\n",
      "      If (feature 153 in {1.0})\n",
      "       If (feature 60 in {1.0})\n",
      "        If (feature 106 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 106 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 60 not in {1.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 153 not in {1.0})\n",
      "       Predict: 1.0\n",
      "    Else (feature 24 not in {1.0})\n",
      "     If (feature 43 in {1.0})\n",
      "      If (feature 19 in {1.0})\n",
      "       If (feature 60 in {1.0})\n",
      "        Predict: 1.0\n",
      "       Else (feature 60 not in {1.0})\n",
      "        If (feature 115 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 115 not in {1.0})\n",
      "         Predict: 1.0\n",
      "      Else (feature 19 not in {1.0})\n",
      "       If (feature 118 in {1.0})\n",
      "        If (feature 9 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 9 not in {1.0})\n",
      "         Predict: 0.0\n",
      "       Else (feature 118 not in {1.0})\n",
      "        If (feature 108 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 108 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 43 not in {1.0})\n",
      "      If (feature 59 in {1.0})\n",
      "       If (feature 169 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 169 not in {1.0})\n",
      "        If (feature 161 in {0.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 161 not in {0.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 59 not in {1.0})\n",
      "       If (feature 2 <= 0.7233277630270138)\n",
      "        If (feature 0 <= 6.777901014488855)\n",
      "         Predict: 0.0\n",
      "        Else (feature 0 > 6.777901014488855)\n",
      "         Predict: 1.0\n",
      "       Else (feature 2 > 0.7233277630270138)\n",
      "        Predict: 1.0\n",
      "  Tree 5 (weight 1.0):\n",
      "    If (feature 36 in {1.0})\n",
      "     If (feature 5 <= 1.3418406115441774)\n",
      "      If (feature 169 in {1.0})\n",
      "       If (feature 59 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 59 not in {1.0})\n",
      "        If (feature 70 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 70 not in {1.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 169 not in {1.0})\n",
      "       If (feature 58 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 58 not in {1.0})\n",
      "        If (feature 148 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 148 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 5 > 1.3418406115441774)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       If (feature 51 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 51 not in {1.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       Predict: 1.0\n",
      "    Else (feature 36 not in {1.0})\n",
      "     If (feature 121 in {0.0})\n",
      "      If (feature 1 <= 3.403313758174694)\n",
      "       If (feature 160 in {0.0})\n",
      "        If (feature 96 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 96 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 160 not in {0.0})\n",
      "        If (feature 44 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 44 not in {1.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 1 > 3.403313758174694)\n",
      "       Predict: 1.0\n",
      "     Else (feature 121 not in {0.0})\n",
      "      If (feature 21 in {1.0})\n",
      "       If (feature 149 in {1.0})\n",
      "        If (feature 171 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 171 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 149 not in {1.0})\n",
      "        Predict: 0.0\n",
      "      Else (feature 21 not in {1.0})\n",
      "       Predict: 0.0\n",
      "  Tree 6 (weight 1.0):\n",
      "    If (feature 149 in {1.0})\n",
      "     If (feature 151 in {0.0})\n",
      "      If (feature 49 in {1.0})\n",
      "       If (feature 169 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 169 not in {1.0})\n",
      "        If (feature 126 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 126 not in {0.0})\n",
      "         Predict: 1.0\n",
      "      Else (feature 49 not in {1.0})\n",
      "       If (feature 145 in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "       Else (feature 145 not in {1.0})\n",
      "        If (feature 43 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 43 not in {1.0})\n",
      "         Predict: 0.0\n",
      "     Else (feature 151 not in {0.0})\n",
      "      If (feature 160 in {1.0})\n",
      "       If (feature 2 <= 0.7233277630270138)\n",
      "        If (feature 1 <= 3.403313758174694)\n",
      "         Predict: 0.0\n",
      "        Else (feature 1 > 3.403313758174694)\n",
      "         Predict: 1.0\n",
      "       Else (feature 2 > 0.7233277630270138)\n",
      "        Predict: 1.0\n",
      "      Else (feature 160 not in {1.0})\n",
      "       Predict: 1.0\n",
      "    Else (feature 149 not in {1.0})\n",
      "     If (feature 59 in {1.0})\n",
      "      If (feature 67 in {1.0})\n",
      "       If (feature 34 in {1.0})\n",
      "        Predict: 0.0\n",
      "       Else (feature 34 not in {1.0})\n",
      "        If (feature 36 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 36 not in {1.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 67 not in {1.0})\n",
      "       If (feature 8 in {1.0})\n",
      "        If (feature 1 <= 3.403313758174694)\n",
      "         Predict: 0.0\n",
      "        Else (feature 1 > 3.403313758174694)\n",
      "         Predict: 1.0\n",
      "       Else (feature 8 not in {1.0})\n",
      "        Predict: 0.0\n",
      "     Else (feature 59 not in {1.0})\n",
      "      If (feature 42 in {1.0})\n",
      "       If (feature 21 in {0.0})\n",
      "        If (feature 23 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 23 not in {1.0})\n",
      "         Predict: 0.0\n",
      "       Else (feature 21 not in {0.0})\n",
      "        Predict: 0.0\n",
      "      Else (feature 42 not in {1.0})\n",
      "       If (feature 1 <= 3.403313758174694)\n",
      "        Predict: 0.0\n",
      "       Else (feature 1 > 3.403313758174694)\n",
      "        Predict: 1.0\n",
      "  Tree 7 (weight 1.0):\n",
      "    If (feature 49 in {1.0})\n",
      "     If (feature 147 in {0.0})\n",
      "      If (feature 148 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 148 not in {1.0})\n",
      "       If (feature 4 <= -0.009973080036487225)\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "       Else (feature 4 > -0.009973080036487225)\n",
      "        If (feature 160 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 160 not in {1.0})\n",
      "         Predict: 1.0\n",
      "     Else (feature 147 not in {0.0})\n",
      "      If (feature 1 <= 3.403313758174694)\n",
      "       Predict: 0.0\n",
      "      Else (feature 1 > 3.403313758174694)\n",
      "       Predict: 1.0\n",
      "    Else (feature 49 not in {1.0})\n",
      "     If (feature 9 in {1.0})\n",
      "      If (feature 60 in {1.0})\n",
      "       If (feature 124 in {1.0})\n",
      "        If (feature 4 <= -0.009973080036487225)\n",
      "         Predict: 1.0\n",
      "        Else (feature 4 > -0.009973080036487225)\n",
      "         Predict: 0.0\n",
      "       Else (feature 124 not in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "      Else (feature 60 not in {1.0})\n",
      "       If (feature 6 <= 0.1070547951588865)\n",
      "        Predict: 0.0\n",
      "       Else (feature 6 > 0.1070547951588865)\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "     Else (feature 9 not in {1.0})\n",
      "      If (feature 106 in {0.0})\n",
      "       If (feature 44 in {1.0})\n",
      "        If (feature 89 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 89 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 44 not in {1.0})\n",
      "        If (feature 43 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 43 not in {1.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 106 not in {0.0})\n",
      "       If (feature 2 <= 0.7233277630270138)\n",
      "        Predict: 0.0\n",
      "       Else (feature 2 > 0.7233277630270138)\n",
      "        Predict: 1.0\n",
      "  Tree 8 (weight 1.0):\n",
      "    If (feature 138 in {1.0})\n",
      "     If (feature 48 in {0.0})\n",
      "      If (feature 23 in {1.0})\n",
      "       If (feature 171 in {1.0})\n",
      "        If (feature 50 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 50 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 171 not in {1.0})\n",
      "        If (feature 149 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 149 not in {0.0})\n",
      "         Predict: 1.0\n",
      "      Else (feature 23 not in {1.0})\n",
      "       If (feature 150 in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "       Else (feature 150 not in {1.0})\n",
      "        Predict: 0.0\n",
      "     Else (feature 48 not in {0.0})\n",
      "      If (feature 43 in {1.0})\n",
      "       If (feature 152 in {1.0})\n",
      "        If (feature 122 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 122 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 152 not in {1.0})\n",
      "        Predict: 1.0\n",
      "      Else (feature 43 not in {1.0})\n",
      "       If (feature 24 in {1.0})\n",
      "        Predict: 1.0\n",
      "       Else (feature 24 not in {1.0})\n",
      "        Predict: 0.0\n",
      "    Else (feature 138 not in {1.0})\n",
      "     If (feature 155 in {1.0})\n",
      "      If (feature 2 <= 0.7233277630270138)\n",
      "       Predict: 0.0\n",
      "      Else (feature 2 > 0.7233277630270138)\n",
      "       Predict: 1.0\n",
      "     Else (feature 155 not in {1.0})\n",
      "      If (feature 44 in {1.0})\n",
      "       Predict: 1.0\n",
      "      Else (feature 44 not in {1.0})\n",
      "       If (feature 58 in {1.0})\n",
      "        If (feature 22 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 22 not in {1.0})\n",
      "         Predict: 0.0\n",
      "       Else (feature 58 not in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "  Tree 9 (weight 1.0):\n",
      "    If (feature 49 in {1.0})\n",
      "     If (feature 147 in {0.0})\n",
      "      If (feature 154 in {1.0})\n",
      "       If (feature 6 <= 0.1070547951588865)\n",
      "        If (feature 94 in {0.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 94 not in {0.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 6 > 0.1070547951588865)\n",
      "        If (feature 106 in {0.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 106 not in {0.0})\n",
      "         Predict: 0.0\n",
      "      Else (feature 154 not in {1.0})\n",
      "       If (feature 59 in {1.0})\n",
      "        Predict: 1.0\n",
      "       Else (feature 59 not in {1.0})\n",
      "        Predict: 0.0\n",
      "     Else (feature 147 not in {0.0})\n",
      "      If (feature 42 in {1.0})\n",
      "       Predict: 0.0\n",
      "      Else (feature 42 not in {1.0})\n",
      "       If (feature 0 <= 6.777901014488855)\n",
      "        Predict: 0.0\n",
      "       Else (feature 0 > 6.777901014488855)\n",
      "        Predict: 1.0\n",
      "    Else (feature 49 not in {1.0})\n",
      "     If (feature 43 in {1.0})\n",
      "      If (feature 4 <= -0.009973080036487225)\n",
      "       If (feature 6 <= 0.1070547951588865)\n",
      "        If (feature 27 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 27 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 6 > 0.1070547951588865)\n",
      "        If (feature 34 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 34 not in {1.0})\n",
      "         Predict: 1.0\n",
      "      Else (feature 4 > -0.009973080036487225)\n",
      "       Predict: 1.0\n",
      "     Else (feature 43 not in {1.0})\n",
      "      If (feature 8 in {1.0})\n",
      "       If (feature 36 in {1.0})\n",
      "        If (feature 90 in {1.0})\n",
      "         Predict: 0.0\n",
      "        Else (feature 90 not in {1.0})\n",
      "         Predict: 1.0\n",
      "       Else (feature 36 not in {1.0})\n",
      "        If (feature 2 <= 0.7233277630270138)\n",
      "         Predict: 0.0\n",
      "        Else (feature 2 > 0.7233277630270138)\n",
      "         Predict: 1.0\n",
      "      Else (feature 8 not in {1.0})\n",
      "       If (feature 151 in {0.0})\n",
      "        If (feature 0 <= 6.777901014488855)\n",
      "         Predict: 0.0\n",
      "        Else (feature 0 > 6.777901014488855)\n",
      "         Predict: 1.0\n",
      "       Else (feature 151 not in {0.0})\n",
      "        If (feature 44 in {1.0})\n",
      "         Predict: 1.0\n",
      "        Else (feature 44 not in {1.0})\n",
      "         Predict: 0.0\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rfModel = rf.fit(testData)\n",
    "print(rfModel.toDebugString)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.5.4 Gradient boosted tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GBTClassificationModel (uid=GBTClassifier_5dc7a7bfd9e7) with 10 trees\n",
      "  Tree 0 (weight 1.0):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       Predict: -1.0\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       Predict: 1.0\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      Predict: 1.0\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     Predict: 1.0\n",
      "  Tree 1 (weight 0.1):\n",
      "    If (feature 2 <= -0.2554261390912864)\n",
      "     If (feature 1 <= -0.07193706104772389)\n",
      "      If (feature 0 <= -0.03668591112962441)\n",
      "       If (feature 5 <= 1.3418406115441774)\n",
      "        If (feature 48 in {0.0})\n",
      "         Predict: -0.47681168808860885\n",
      "        Else (feature 48 not in {0.0})\n",
      "         Predict: -0.4768116880882528\n",
      "       Else (feature 5 > 1.3418406115441774)\n",
      "        If (feature 67 in {0.0})\n",
      "         Predict: -0.4768116880884589\n",
      "        Else (feature 67 not in {0.0})\n",
      "         Predict: -0.4768116880884531\n",
      "      Else (feature 0 > -0.03668591112962441)\n",
      "       Predict: 0.47681168808847285\n",
      "     Else (feature 1 > -0.07193706104772389)\n",
      "      If (feature 17 in {0.0})\n",
      "       If (feature 138 in {1.0})\n",
      "        If (feature 37 in {1.0})\n",
      "         Predict: 0.4768116880884702\n",
      "        Else (feature 37 not in {1.0})\n",
      "         Predict: 0.4768116880884717\n",
      "       Else (feature 138 not in {1.0})\n",
      "        If (feature 65 in {1.0})\n",
      "         Predict: 0.47681168808847013\n",
      "        Else (feature 65 not in {1.0})\n",
      "         Predict: 0.476811688088472\n",
      "      Else (feature 17 not in {0.0})\n",
      "       If (feature 138 in {0.0})\n",
      "        If (feature 13 in {1.0})\n",
      "         Predict: 0.4768116880884702\n",
      "        Else (feature 13 not in {1.0})\n",
      "         Predict: 0.4768116880884717\n",
      "       Else (feature 138 not in {0.0})\n",
      "        If (feature 123 in {1.0})\n",
      "         Predict: 0.4768116880884702\n",
      "        Else (feature 123 not in {1.0})\n",
      "         Predict: 0.4768116880884718\n",
      "    Else (feature 2 > -0.2554261390912864)\n",
      "     If (feature 83 in {1.0})\n",
      "      Predict: 0.47681168808847024\n",
      "     Else (feature 83 not in {1.0})\n",
      "      Predict: 0.47681168808860225\n",
      "  Tree 2 (weight 0.1):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       If (feature 171 in {0.0})\n",
      "        If (feature 161 in {1.0})\n",
      "         Predict: -0.4381935810429489\n",
      "        Else (feature 161 not in {1.0})\n",
      "         Predict: -0.4381935810427036\n",
      "       Else (feature 171 not in {0.0})\n",
      "        Predict: -0.43819358104283\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       If (feature 70 in {1.0})\n",
      "        If (feature 54 in {0.0})\n",
      "         Predict: 0.43819358104272044\n",
      "        Else (feature 54 not in {0.0})\n",
      "         Predict: 0.4381935810427205\n",
      "       Else (feature 70 not in {1.0})\n",
      "        Predict: 0.43819358104272144\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      If (feature 17 in {0.0})\n",
      "       If (feature 111 in {1.0})\n",
      "        Predict: 0.4381935810427206\n",
      "       Else (feature 111 not in {1.0})\n",
      "        Predict: 0.4381935810427235\n",
      "      Else (feature 17 not in {0.0})\n",
      "       If (feature 39 in {1.0})\n",
      "        Predict: 0.43819358104272044\n",
      "       Else (feature 39 not in {1.0})\n",
      "        Predict: 0.4381935810427235\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     If (feature 147 in {0.0})\n",
      "      If (feature 84 in {0.0})\n",
      "       Predict: 0.43819358104270406\n",
      "      Else (feature 84 not in {0.0})\n",
      "       Predict: 0.43819358104271017\n",
      "     Else (feature 147 not in {0.0})\n",
      "      Predict: 0.4381935810428038\n",
      "  Tree 3 (weight 0.1):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       Predict: -0.4051496802846631\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       If (feature 5 <= 1.3418406115441774)\n",
      "        Predict: 0.40514968028459747\n",
      "       Else (feature 5 > 1.3418406115441774)\n",
      "        If (feature 95 in {1.0})\n",
      "         Predict: 0.4051496802845981\n",
      "        Else (feature 95 not in {1.0})\n",
      "         Predict: 0.40514968028459813\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      If (feature 138 in {1.0})\n",
      "       If (feature 13 in {1.0})\n",
      "        Predict: 0.405149680284598\n",
      "       Else (feature 13 not in {1.0})\n",
      "        Predict: 0.4051496802846005\n",
      "      Else (feature 138 not in {1.0})\n",
      "       Predict: 0.4051496802846008\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     If (feature 86 in {1.0})\n",
      "      Predict: 0.40514968028458576\n",
      "     Else (feature 86 not in {1.0})\n",
      "      Predict: 0.40514968028464415\n",
      "  Tree 4 (weight 0.1):\n",
      "    If (feature 2 <= -0.2554261390912864)\n",
      "     If (feature 1 <= -0.07193706104772389)\n",
      "      If (feature 0 <= -0.03668591112962441)\n",
      "       If (feature 57 in {1.0})\n",
      "        If (feature 11 in {0.0})\n",
      "         Predict: -0.37658413183538814\n",
      "        Else (feature 11 not in {0.0})\n",
      "         Predict: -0.37658413183528583\n",
      "       Else (feature 57 not in {1.0})\n",
      "        If (feature 87 in {1.0})\n",
      "         Predict: -0.3765841318353918\n",
      "        Else (feature 87 not in {1.0})\n",
      "         Predict: -0.37658413183527856\n",
      "      Else (feature 0 > -0.03668591112962441)\n",
      "       Predict: 0.3765841318353042\n",
      "     Else (feature 1 > -0.07193706104772389)\n",
      "      If (feature 147 in {0.0})\n",
      "       If (feature 83 in {1.0})\n",
      "        Predict: 0.3765841318352986\n",
      "       Else (feature 83 not in {1.0})\n",
      "        If (feature 91 in {1.0})\n",
      "         Predict: 0.37658413183529865\n",
      "        Else (feature 91 not in {1.0})\n",
      "         Predict: 0.37658413183530065\n",
      "      Else (feature 147 not in {0.0})\n",
      "       Predict: 0.3765841318353007\n",
      "    Else (feature 2 > -0.2554261390912864)\n",
      "     Predict: 0.3765841318353693\n",
      "  Tree 5 (weight 0.1):\n",
      "    If (feature 2 <= -0.2554261390912864)\n",
      "     If (feature 1 <= -0.07193706104772389)\n",
      "      If (feature 0 <= -0.03668591112962441)\n",
      "       Predict: -0.3516647895811837\n",
      "      Else (feature 0 > -0.03668591112962441)\n",
      "       Predict: 0.35166478958100844\n",
      "     Else (feature 1 > -0.07193706104772389)\n",
      "      If (feature 83 in {1.0})\n",
      "       Predict: 0.35166478958100983\n",
      "      Else (feature 83 not in {1.0})\n",
      "       Predict: 0.3516647895810165\n",
      "    Else (feature 2 > -0.2554261390912864)\n",
      "     If (feature 5 <= 1.3418406115441774)\n",
      "      If (feature 17 in {1.0})\n",
      "       If (feature 148 in {1.0})\n",
      "        If (feature 47 in {1.0})\n",
      "         Predict: 0.3516647895809951\n",
      "        Else (feature 47 not in {1.0})\n",
      "         Predict: 0.35166478958100045\n",
      "       Else (feature 148 not in {1.0})\n",
      "        If (feature 42 in {1.0})\n",
      "         Predict: 0.35166478958099473\n",
      "        Else (feature 42 not in {1.0})\n",
      "         Predict: 0.35166478958101866\n",
      "      Else (feature 17 not in {1.0})\n",
      "       If (feature 21 in {1.0})\n",
      "        Predict: 0.35166478958099273\n",
      "       Else (feature 21 not in {1.0})\n",
      "        If (feature 148 in {1.0})\n",
      "         Predict: 0.3516647895809967\n",
      "        Else (feature 148 not in {1.0})\n",
      "         Predict: 0.3516647895810213\n",
      "     Else (feature 5 > 1.3418406115441774)\n",
      "      If (feature 3 <= -0.450303698596878)\n",
      "       If (feature 169 in {0.0})\n",
      "        If (feature 92 in {1.0})\n",
      "         Predict: 0.3516647895809889\n",
      "        Else (feature 92 not in {1.0})\n",
      "         Predict: 0.35166478958099134\n",
      "       Else (feature 169 not in {0.0})\n",
      "        If (feature 108 in {1.0})\n",
      "         Predict: 0.35166478958098923\n",
      "        Else (feature 108 not in {1.0})\n",
      "         Predict: 0.35166478958099145\n",
      "      Else (feature 3 > -0.450303698596878)\n",
      "       If (feature 106 in {0.0})\n",
      "        If (feature 68 in {1.0})\n",
      "         Predict: 0.3516647895809894\n",
      "        Else (feature 68 not in {1.0})\n",
      "         Predict: 0.3516647895809901\n",
      "       Else (feature 106 not in {0.0})\n",
      "        If (feature 147 in {1.0})\n",
      "         Predict: 0.35166478958099046\n",
      "        Else (feature 147 not in {1.0})\n",
      "         Predict: 0.351664789580995\n",
      "  Tree 6 (weight 0.1):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       If (feature 4 <= -0.009973080036487225)\n",
      "        If (feature 70 in {0.0})\n",
      "         Predict: -0.3297498465553543\n",
      "        Else (feature 70 not in {0.0})\n",
      "         Predict: -0.32974984655528156\n",
      "       Else (feature 4 > -0.009973080036487225)\n",
      "        If (feature 109 in {0.0})\n",
      "         Predict: -0.3297498465553619\n",
      "        Else (feature 109 not in {0.0})\n",
      "         Predict: -0.32974984655527895\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       If (feature 53 in {1.0})\n",
      "        Predict: 0.329749846555299\n",
      "       Else (feature 53 not in {1.0})\n",
      "        Predict: 0.3297498465553005\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      If (feature 147 in {0.0})\n",
      "       Predict: 0.3297498465553012\n",
      "      Else (feature 147 not in {0.0})\n",
      "       If (feature 119 in {1.0})\n",
      "        Predict: 0.32974984655529843\n",
      "       Else (feature 119 not in {1.0})\n",
      "        Predict: 0.3297498465553014\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     If (feature 172 in {1.0})\n",
      "      Predict: 0.3297498465553211\n",
      "     Else (feature 172 not in {1.0})\n",
      "      If (feature 47 in {1.0})\n",
      "       If (feature 25 in {0.0})\n",
      "        If (feature 10 in {0.0})\n",
      "         Predict: 0.32974984655528194\n",
      "        Else (feature 10 not in {0.0})\n",
      "         Predict: 0.32974984655528217\n",
      "       Else (feature 25 not in {0.0})\n",
      "        Predict: 0.32974984655528816\n",
      "      Else (feature 47 not in {1.0})\n",
      "       Predict: 0.3297498465553503\n",
      "  Tree 7 (weight 0.1):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       If (feature 4 <= -0.009973080036487225)\n",
      "        If (feature 59 in {0.0})\n",
      "         Predict: -0.3103372455198011\n",
      "        Else (feature 59 not in {0.0})\n",
      "         Predict: -0.31033724551979475\n",
      "       Else (feature 4 > -0.009973080036487225)\n",
      "        If (feature 11 in {0.0})\n",
      "         Predict: -0.31033724551980124\n",
      "        Else (feature 11 not in {0.0})\n",
      "         Predict: -0.31033724551979763\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       Predict: 0.3103372455197969\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      If (feature 157 in {1.0})\n",
      "       Predict: 0.31033724551979464\n",
      "      Else (feature 157 not in {1.0})\n",
      "       If (feature 71 in {1.0})\n",
      "        Predict: 0.31033724551979464\n",
      "       Else (feature 71 not in {1.0})\n",
      "        Predict: 0.3103372455197995\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     If (feature 147 in {1.0})\n",
      "      If (feature 40 in {1.0})\n",
      "       Predict: 0.31033724551977354\n",
      "      Else (feature 40 not in {1.0})\n",
      "       If (feature 47 in {1.0})\n",
      "        Predict: 0.310337245519778\n",
      "       Else (feature 47 not in {1.0})\n",
      "        Predict: 0.3103372455197966\n",
      "     Else (feature 147 not in {1.0})\n",
      "      If (feature 105 in {1.0})\n",
      "       If (feature 169 in {0.0})\n",
      "        Predict: 0.310337245519779\n",
      "       Else (feature 169 not in {0.0})\n",
      "        Predict: 0.3103372455197794\n",
      "      Else (feature 105 not in {1.0})\n",
      "       Predict: 0.31033724551979885\n",
      "  Tree 8 (weight 0.1):\n",
      "    If (feature 2 <= 0.7233277630270138)\n",
      "     If (feature 1 <= 3.403313758174694)\n",
      "      If (feature 0 <= 6.777901014488855)\n",
      "       If (feature 57 in {0.0})\n",
      "        Predict: -0.2930291649126393\n",
      "       Else (feature 57 not in {0.0})\n",
      "        If (feature 27 in {0.0})\n",
      "         Predict: -0.2930291649126371\n",
      "        Else (feature 27 not in {0.0})\n",
      "         Predict: -0.2930291649126957\n",
      "      Else (feature 0 > 6.777901014488855)\n",
      "       If (feature 5 <= 1.3418406115441774)\n",
      "        If (feature 15 in {0.0})\n",
      "         Predict: 0.29302916491254294\n",
      "        Else (feature 15 not in {0.0})\n",
      "         Predict: 0.293029164912543\n",
      "       Else (feature 5 > 1.3418406115441774)\n",
      "        Predict: 0.2930291649125431\n",
      "     Else (feature 1 > 3.403313758174694)\n",
      "      Predict: 0.2930291649125445\n",
      "    Else (feature 2 > 0.7233277630270138)\n",
      "     If (feature 157 in {1.0})\n",
      "      Predict: 0.29302916491252823\n",
      "     Else (feature 157 not in {1.0})\n",
      "      Predict: 0.29302916491259134\n",
      "  Tree 9 (weight 0.1):\n",
      "    If (feature 2 <= -0.2554261390912864)\n",
      "     If (feature 1 <= -0.07193706104772389)\n",
      "      If (feature 0 <= -0.03668591112962441)\n",
      "       Predict: -0.27750666438388916\n",
      "      Else (feature 0 > -0.03668591112962441)\n",
      "       If (feature 17 in {0.0})\n",
      "        If (feature 162 in {1.0})\n",
      "         Predict: 0.2775066643835821\n",
      "        Else (feature 162 not in {1.0})\n",
      "         Predict: 0.2775066643835827\n",
      "       Else (feature 17 not in {0.0})\n",
      "        Predict: 0.2775066643835826\n",
      "     Else (feature 1 > -0.07193706104772389)\n",
      "      If (feature 31 in {1.0})\n",
      "       If (feature 15 in {0.0})\n",
      "        If (feature 59 in {1.0})\n",
      "         Predict: 0.27750666438358146\n",
      "        Else (feature 59 not in {1.0})\n",
      "         Predict: 0.2775066643835816\n",
      "       Else (feature 15 not in {0.0})\n",
      "        Predict: 0.2775066643835817\n",
      "      Else (feature 31 not in {1.0})\n",
      "       If (feature 104 in {1.0})\n",
      "        If (feature 59 in {1.0})\n",
      "         Predict: 0.27750666438358146\n",
      "        Else (feature 59 not in {1.0})\n",
      "         Predict: 0.2775066643835815\n",
      "       Else (feature 104 not in {1.0})\n",
      "        If (feature 120 in {1.0})\n",
      "         Predict: 0.27750666438358146\n",
      "        Else (feature 120 not in {1.0})\n",
      "         Predict: 0.27750666438358623\n",
      "    Else (feature 2 > -0.2554261390912864)\n",
      "     If (feature 5 <= -0.5923165974980282)\n",
      "      If (feature 138 in {0.0})\n",
      "       If (feature 19 in {1.0})\n",
      "        Predict: 0.2775066643835649\n",
      "       Else (feature 19 not in {1.0})\n",
      "        Predict: 0.277506664383565\n",
      "      Else (feature 138 not in {0.0})\n",
      "       If (feature 7 in {1.0})\n",
      "        If (feature 77 in {1.0})\n",
      "         Predict: 0.27750666438356125\n",
      "        Else (feature 77 not in {1.0})\n",
      "         Predict: 0.2775066643835633\n",
      "       Else (feature 7 not in {1.0})\n",
      "        Predict: 0.2775066643835651\n",
      "     Else (feature 5 > -0.5923165974980282)\n",
      "      If (feature 147 in {1.0})\n",
      "       If (feature 19 in {1.0})\n",
      "        If (feature 15 in {1.0})\n",
      "         Predict: 0.2775066643835645\n",
      "        Else (feature 15 not in {1.0})\n",
      "         Predict: 0.27750666438356475\n",
      "       Else (feature 19 not in {1.0})\n",
      "        If (feature 113 in {1.0})\n",
      "         Predict: 0.2775066643835626\n",
      "        Else (feature 113 not in {1.0})\n",
      "         Predict: 0.27750666438356497\n",
      "      Else (feature 147 not in {1.0})\n",
      "       If (feature 150 in {1.0})\n",
      "        If (feature 38 in {0.0})\n",
      "         Predict: 0.2775066643835656\n",
      "        Else (feature 38 not in {0.0})\n",
      "         Predict: 0.27750666438356575\n",
      "       Else (feature 150 not in {1.0})\n",
      "        If (feature 98 in {1.0})\n",
      "         Predict: 0.2775066643835655\n",
      "        Else (feature 98 not in {1.0})\n",
      "         Predict: 0.27750666438357213\n",
      "\n"
     ]
    }
   ],
   "source": [
    "gbtModel = gbt.fit(trainData)\n",
    "print(gbtModel.toDebugString)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.6 Binary classification evaluator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DECISION TREE Area under ROC: 0.6030471193727109\n",
      "RANDOM FOREST Area under ROC: 0.8635932161985542\n",
      "GBT Area under ROC: 0.8981134828142829\n",
      "LOGISTIC REGRESSION Area under ROC: 0.9139872701825695\n",
      "NAIVE BAYES Area under ROC: 0.7712670435026981\n"
     ]
    }
   ],
   "source": [
    "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
    "\n",
    "# DECISION TREE\n",
    "dtEvaluator = BinaryClassificationEvaluator(metricName=\"areaUnderROC\", labelCol=\"label\", rawPredictionCol=\"DTRawPrediction\")\n",
    "print(\"DECISION TREE Area under ROC: \" + str(dtEvaluator.evaluate(predictions)))\n",
    "\n",
    "#RANDOM FOREST\n",
    "rfEvaluator = BinaryClassificationEvaluator(metricName=\"areaUnderROC\", labelCol=\"label\", rawPredictionCol=\"RFRawPrediction\")\n",
    "print(\"RANDOM FOREST Area under ROC: \" + str(rfEvaluator.evaluate(predictions)))\n",
    "\n",
    "#GBT\n",
    "gbtEvaluator = BinaryClassificationEvaluator(metricName=\"areaUnderROC\", labelCol=\"label\", rawPredictionCol=\"rawPrediction\")\n",
    "print(\"GBT Area under ROC: \" + str(gbtEvaluator.evaluate(predictions)))\n",
    "\n",
    "#LOGISTIC REGRESSION\n",
    "blrEvaluator = BinaryClassificationEvaluator(metricName=\"areaUnderROC\", labelCol=\"label\", rawPredictionCol=\"LRRawPrediction\")\n",
    "print(\"LOGISTIC REGRESSION Area under ROC: \" + str(blrEvaluator.evaluate(predictions)))\n",
    "\n",
    "#NAIVE BAYES\n",
    "nbEvaluator = BinaryClassificationEvaluator(metricName=\"areaUnderROC\", labelCol=\"label\", rawPredictionCol=\"NBprediction\")\n",
    "print(\"NAIVE BAYES Area under ROC: \" + str(nbEvaluator.evaluate(predictions)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3.6.1 Logit metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 0.8792841357248385\n",
      "TPR (Recall/Sensitivity): 0.8792841357248385\n"
     ]
    }
   ],
   "source": [
    "# create summary for logit and calculate Accuracy and TPR\n",
    "trainingSummary = lrModel.summary\n",
    "\n",
    "accuracy = trainingSummary.accuracy #correctly predicted TP + TN\n",
    "truePositiveRate = trainingSummary.weightedTruePositiveRate #correctly predicted TN\n",
    "print(\"Accuracy: %s\\nTPR (Recall/Sensitivity): %s\"\n",
    "      % (accuracy, truePositiveRate))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3.7 Cross validation and other metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### (Not used)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "from pyspark.ml.tuning import CrossValidator, ParamGridBuilder\n",
    "from pyspark.ml.evaluation import BinaryClassificationEvaluator\n",
    "\n",
    "blrPipeline = Pipeline(stages = [lr])\n",
    "\n",
    "blrParamGrid = ParamGridBuilder() \\\n",
    "    .addGrid(lr.elasticNetParam, [0.0, 0.1, 0.5, 0.9, 1.0])\\\n",
    "    .addGrid(lr.regParam, [0.0, 0.5, 0.9])\\\n",
    "    .addGrid(lr.maxIter, [50]).build()\n",
    "    \n",
    "blrCrossValidator = CrossValidator(estimator=blrPipeline,\n",
    "                          estimatorParamMaps=blrParamGrid,\n",
    "                          evaluator=blrEvaluator,\n",
    "                          numFolds=2,\n",
    "                         parallelism=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "blrCrossValidatorModel = blrCrossValidator.fit(modelData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['elasticNetParam', 'regParam', 'maxIter']\n"
     ]
    }
   ],
   "source": [
    "l1 = len(blrParamGrid[0].keys()) \n",
    "paramName = []\n",
    "\n",
    "for i in range(0,l1):\n",
    "  paramName.append(list(blrParamGrid[0].keys())[i].name)\n",
    "\n",
    "print(paramName)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[dict_values([0.0, 0.0, 50]), dict_values([0.0, 0.5, 50]), dict_values([0.0, 0.9, 50]), dict_values([0.1, 0.0, 50]), dict_values([0.1, 0.5, 50]), dict_values([0.1, 0.9, 50]), dict_values([0.5, 0.0, 50]), dict_values([0.5, 0.5, 50]), dict_values([0.5, 0.9, 50]), dict_values([0.9, 0.0, 50]), dict_values([0.9, 0.5, 50]), dict_values([0.9, 0.9, 50]), dict_values([1.0, 0.0, 50]), dict_values([1.0, 0.5, 50]), dict_values([1.0, 0.9, 50])]\n"
     ]
    }
   ],
   "source": [
    "l2 = len(blrParamGrid)\n",
    "paramValues = []\n",
    "\n",
    "for i in range(0,l2):\n",
    "  paramValues.append(blrParamGrid[i].values())\n",
    "\n",
    "print(paramValues)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>elasticNetParam</th>\n",
       "      <th>regParam</th>\n",
       "      <th>maxIter</th>\n",
       "      <th>areaUnderROC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0.913691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>50</td>\n",
       "      <td>0.885230</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>0.9</td>\n",
       "      <td>50</td>\n",
       "      <td>0.877250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0.913691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.5</td>\n",
       "      <td>50</td>\n",
       "      <td>0.850549</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.1</td>\n",
       "      <td>0.9</td>\n",
       "      <td>50</td>\n",
       "      <td>0.784078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0.913691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.5</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.5</td>\n",
       "      <td>0.9</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0.913691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.5</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.9</td>\n",
       "      <td>0.9</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>50</td>\n",
       "      <td>0.913691</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.9</td>\n",
       "      <td>50</td>\n",
       "      <td>0.500000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    elasticNetParam  regParam  maxIter  areaUnderROC\n",
       "0               0.0       0.0       50      0.913691\n",
       "1               0.0       0.5       50      0.885230\n",
       "2               0.0       0.9       50      0.877250\n",
       "3               0.1       0.0       50      0.913691\n",
       "4               0.1       0.5       50      0.850549\n",
       "5               0.1       0.9       50      0.784078\n",
       "6               0.5       0.0       50      0.913691\n",
       "7               0.5       0.5       50      0.500000\n",
       "8               0.5       0.9       50      0.500000\n",
       "9               0.9       0.0       50      0.913691\n",
       "10              0.9       0.5       50      0.500000\n",
       "11              0.9       0.9       50      0.500000\n",
       "12              1.0       0.0       50      0.913691\n",
       "13              1.0       0.5       50      0.500000\n",
       "14              1.0       0.9       50      0.500000"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "allBlrModels = pd.DataFrame.from_records(paramValues, columns =paramName )\n",
    "\n",
    "metricName = blrEvaluator.getMetricName()\n",
    "allBlrModels.loc[:,metricName] = blrCrossValidatorModel.avgMetrics\n",
    "\n",
    "allBlrModels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DataFrame[label: int, features: vector, LRRawPrediction: vector, LRProbability: vector, LRprediction: double]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bestBlrModel = blrCrossValidatorModel.bestModel\n",
    "bestBlrModelI = bestBlrModel.stages[-1]\n",
    "bestBlrModel.transform(testData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-1.4210933056043655"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "intercept = bestBlrModelI.intercept\n",
    "intercept"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "FPR = list(chain(*bestBlrModel.stages[-1].summary.roc.select(\"FPR\").toPandas().values))\n",
    "TPR = list(chain(*bestBlrModel.stages[-1].summary.roc.select(\"TPR\").toPandas().values))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plotROC(x, y):\n",
    "    plt.scatter(x, y, color = \"#00800D\")\n",
    "    #plt.axis(\"off\")\n",
    "    plt.title(r'ROC', fontsize=15, weight = 'bold')\n",
    "    plt.xlabel(\"False Positive rate\", color = 'black',  fontsize=8, weight = 'bold')\n",
    "    plt.ylabel(\"True positive rate\", color = 'black',  fontsize=8, weight = 'bold')\n",
    "    plt.grid(True, color = 'lightgrey')\n",
    "    plt.axis([0, 1, 0, 1])\n",
    "    plt.viridis()\n",
    "\n",
    "    x1=[0,1]\n",
    "    y1=[0,1]\n",
    "    plt.plot(x1,y1)\n",
    "    plt.show()\n",
    "    return; \n",
    "\n",
    "plotROC(FPR, TPR)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "blrResults = bestBlrModel.transform(modelData)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "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>label</th>\n",
       "      <th>1.0</th>\n",
       "      <th>0.0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>44442</td>\n",
       "      <td>42087</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>14500</td>\n",
       "      <td>367264</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   label    1.0     0.0\n",
       "0      1  44442   42087\n",
       "1      0  14500  367264"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "blrResultMatrix = blrResults\\\n",
    "                  .orderBy(col(\"label\").desc())\\\n",
    "                  .groupBy(\"label\")\\\n",
    "                  .pivot(\"lrPrediction\" , [1.0, 0.0])\\\n",
    "                  .count()\n",
    "        \n",
    "display(blrResultMatrix.toPandas())"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}