model accuracy.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 596,
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import glob\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import flat_table\n",
    "import random\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import metrics\n",
    "\n",
    "\n",
    "path = './data/road-game-default-rtdb-export_prev_prev.json'\n",
    "branchs = ['all-games','complete-games']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 597,
   "metadata": {
    "scrolled": true
   },
   "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>index</th>\n",
       "      <th>additional_comments</th>\n",
       "      <th>aggressively_value</th>\n",
       "      <th>behavior</th>\n",
       "      <th>birthYear</th>\n",
       "      <th>blueScore</th>\n",
       "      <th>computer_value</th>\n",
       "      <th>drivinglicense</th>\n",
       "      <th>education</th>\n",
       "      <th>gender</th>\n",
       "      <th>generously_value</th>\n",
       "      <th>log</th>\n",
       "      <th>press-the-read-instructions-button</th>\n",
       "      <th>redScore</th>\n",
       "      <th>steps</th>\n",
       "      <th>win</th>\n",
       "      <th>wisely_value</th>\n",
       "      <th>predictable_value</th>\n",
       "      <th>answer-the-quiz</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-MZiWIwdnrn8TJ034z2w</td>\n",
       "      <td></td>\n",
       "      <td>5.00</td>\n",
       "      <td>carefulBehavior</td>\n",
       "      <td>1961</td>\n",
       "      <td>20.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>valid</td>\n",
       "      <td>BA</td>\n",
       "      <td>male</td>\n",
       "      <td>6.00</td>\n",
       "      <td>[None, {'blue': 'from: a6, action: left', 'red...</td>\n",
       "      <td>True</td>\n",
       "      <td>23.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>True</td>\n",
       "      <td>1.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-MZiWUxZs8BvWglGhc9r</td>\n",
       "      <td></td>\n",
       "      <td>2.00</td>\n",
       "      <td>carefulBehavior</td>\n",
       "      <td>1982</td>\n",
       "      <td>19.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>valid</td>\n",
       "      <td>BA</td>\n",
       "      <td>female</td>\n",
       "      <td>4.00</td>\n",
       "      <td>[None, {'blue': 'from: a6, action: left', 'red...</td>\n",
       "      <td>True</td>\n",
       "      <td>22.00</td>\n",
       "      <td>8.00</td>\n",
       "      <td>True</td>\n",
       "      <td>4.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-MZiWYgxyToIiCSDQY8S</td>\n",
       "      <td></td>\n",
       "      <td>5.00</td>\n",
       "      <td>aggressiveBehavior</td>\n",
       "      <td>1953</td>\n",
       "      <td>-105.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>valid</td>\n",
       "      <td>BA</td>\n",
       "      <td>female</td>\n",
       "      <td>2.00</td>\n",
       "      <td>[None, {'blue': 'from: a6, action: left', 'red...</td>\n",
       "      <td>True</td>\n",
       "      <td>-105.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>False</td>\n",
       "      <td>6.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-MZiWYz_z_BcQlpReXf_</td>\n",
       "      <td></td>\n",
       "      <td>6.00</td>\n",
       "      <td>aggressiveBehavior</td>\n",
       "      <td>1990</td>\n",
       "      <td>25.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>valid</td>\n",
       "      <td>Master's degree</td>\n",
       "      <td>male</td>\n",
       "      <td>2.00</td>\n",
       "      <td>[None, {'blue': 'from: a6, action: left', 'red...</td>\n",
       "      <td>True</td>\n",
       "      <td>21.00</td>\n",
       "      <td>9.00</td>\n",
       "      <td>True</td>\n",
       "      <td>2.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-MZiWaNZ6n3K3AZDfr-q</td>\n",
       "      <td></td>\n",
       "      <td>6.00</td>\n",
       "      <td>aggressiveBehavior</td>\n",
       "      <td>1953</td>\n",
       "      <td>-105.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>valid</td>\n",
       "      <td>BA</td>\n",
       "      <td>female</td>\n",
       "      <td>2.00</td>\n",
       "      <td>[None, {'blue': 'from: a6, action: left', 'red...</td>\n",
       "      <td>True</td>\n",
       "      <td>-105.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>False</td>\n",
       "      <td>1.00</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>667</th>\n",
       "      <td>-MbcRp52SnKf2_aDNt2a</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>valueIterationBasedBehavior_v2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>668</th>\n",
       "      <td>-MbcURCQaLlJFHGxAvyW</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>valueIterationBasedBehavior_v2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>669</th>\n",
       "      <td>-MbcbjkKh8M10JRrqfwb</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>valueIterationBasedBehavior_v2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>670</th>\n",
       "      <td>-MbcfQpUktm_lYEmxpz2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>valueIterationBasedBehavior_v2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>671</th>\n",
       "      <td>-Mbcz6inzp9NEdJCbK17</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>valueIterationBasedBehavior_v2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>672 rows × 19 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                    index additional_comments  aggressively_value  \\\n",
       "0    -MZiWIwdnrn8TJ034z2w                                    5.00   \n",
       "1    -MZiWUxZs8BvWglGhc9r                                    2.00   \n",
       "2    -MZiWYgxyToIiCSDQY8S                                    5.00   \n",
       "3    -MZiWYz_z_BcQlpReXf_                                    6.00   \n",
       "4    -MZiWaNZ6n3K3AZDfr-q                                    6.00   \n",
       "..                    ...                 ...                 ...   \n",
       "667  -MbcRp52SnKf2_aDNt2a                 NaN                 NaN   \n",
       "668  -MbcURCQaLlJFHGxAvyW                 NaN                 NaN   \n",
       "669  -MbcbjkKh8M10JRrqfwb                 NaN                 NaN   \n",
       "670  -MbcfQpUktm_lYEmxpz2                 NaN                 NaN   \n",
       "671  -Mbcz6inzp9NEdJCbK17                 NaN                 NaN   \n",
       "\n",
       "                           behavior birthYear  blueScore  computer_value  \\\n",
       "0                   carefulBehavior      1961      20.00            5.00   \n",
       "1                   carefulBehavior      1982      19.00            6.00   \n",
       "2                aggressiveBehavior      1953    -105.00            7.00   \n",
       "3                aggressiveBehavior      1990      25.00            7.00   \n",
       "4                aggressiveBehavior      1953    -105.00            6.00   \n",
       "..                              ...       ...        ...             ...   \n",
       "667  valueIterationBasedBehavior_v2       NaN        NaN             NaN   \n",
       "668  valueIterationBasedBehavior_v2       NaN        NaN             NaN   \n",
       "669  valueIterationBasedBehavior_v2       NaN        NaN             NaN   \n",
       "670  valueIterationBasedBehavior_v2       NaN        NaN             NaN   \n",
       "671  valueIterationBasedBehavior_v2       NaN        NaN             NaN   \n",
       "\n",
       "    drivinglicense        education  gender  generously_value  \\\n",
       "0            valid               BA    male              6.00   \n",
       "1            valid               BA  female              4.00   \n",
       "2            valid               BA  female              2.00   \n",
       "3            valid  Master's degree    male              2.00   \n",
       "4            valid               BA  female              2.00   \n",
       "..             ...              ...     ...               ...   \n",
       "667            NaN              NaN     NaN               NaN   \n",
       "668            NaN              NaN     NaN               NaN   \n",
       "669            NaN              NaN     NaN               NaN   \n",
       "670            NaN              NaN     NaN               NaN   \n",
       "671            NaN              NaN     NaN               NaN   \n",
       "\n",
       "                                                   log  \\\n",
       "0    [None, {'blue': 'from: a6, action: left', 'red...   \n",
       "1    [None, {'blue': 'from: a6, action: left', 'red...   \n",
       "2    [None, {'blue': 'from: a6, action: left', 'red...   \n",
       "3    [None, {'blue': 'from: a6, action: left', 'red...   \n",
       "4    [None, {'blue': 'from: a6, action: left', 'red...   \n",
       "..                                                 ...   \n",
       "667                                                NaN   \n",
       "668                                                NaN   \n",
       "669                                                NaN   \n",
       "670                                                NaN   \n",
       "671                                                NaN   \n",
       "\n",
       "    press-the-read-instructions-button  redScore  steps    win  wisely_value  \\\n",
       "0                                 True     23.00   7.00   True          1.00   \n",
       "1                                 True     22.00   8.00   True          4.00   \n",
       "2                                 True   -105.00   5.00  False          6.00   \n",
       "3                                 True     21.00   9.00   True          2.00   \n",
       "4                                 True   -105.00   5.00  False          1.00   \n",
       "..                                 ...       ...    ...    ...           ...   \n",
       "667                                NaN       NaN    NaN    NaN           NaN   \n",
       "668                                NaN       NaN    NaN    NaN           NaN   \n",
       "669                                NaN       NaN    NaN    NaN           NaN   \n",
       "670                                NaN       NaN    NaN    NaN           NaN   \n",
       "671                                NaN       NaN    NaN    NaN           NaN   \n",
       "\n",
       "     predictable_value answer-the-quiz  \n",
       "0                  NaN             NaN  \n",
       "1                  NaN             NaN  \n",
       "2                  NaN             NaN  \n",
       "3                  NaN             NaN  \n",
       "4                  NaN             NaN  \n",
       "..                 ...             ...  \n",
       "667                NaN             NaN  \n",
       "668                NaN             NaN  \n",
       "669                NaN             NaN  \n",
       "670                NaN             NaN  \n",
       "671                NaN             NaN  \n",
       "\n",
       "[672 rows x 19 columns]"
      ]
     },
     "execution_count": 597,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#load data from the json file\n",
    "with open(path) as train_file:\n",
    "    data = json.load(train_file)\n",
    "    \n",
    "branch = branchs[0]\n",
    "df = pd.DataFrame.from_dict(data[branch], orient='index')\n",
    "df.reset_index(level=0, inplace=True)\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 598,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['index', 'additional_comments', 'aggressively_value', 'behavior',\n",
       "       'birthYear', 'blueScore', 'computer_value', 'drivinglicense',\n",
       "       'education', 'gender', 'generously_value', 'log',\n",
       "       'press-the-read-instructions-button', 'redScore', 'steps', 'win',\n",
       "       'wisely_value', 'predictable_value', 'answer-the-quiz'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 598,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Survay results:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "upload Id workers and search for duplicate workers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 599,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = './data/workers' # use your path\n",
    "all_files = glob.glob(path + \"/*.csv\")\n",
    "\n",
    "li = []\n",
    "\n",
    "for filename in all_files:\n",
    "    fromOne = pd.read_csv(filename, index_col=None, header=0)\n",
    "    fromOne['filename'] = filename\n",
    "    li.append(fromOne)\n",
    "\n",
    "workers = pd.concat(li, axis=0, ignore_index=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 600,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['WorkerId'] = workers['WorkerId']\n",
    "df['filename'] = workers['filename'].apply(lambda x: x.split('./data/workers/Batch_')[1].split('_batch_results.csv')[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 601,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "WorkerId\n",
       "A17AF42SNQNH9C    8\n",
       "A1AZLJ3R5FQ9B7    2\n",
       "A1BQI9JKLVCIL0    2\n",
       "A1GC6CN3F0IHWN    2\n",
       "A1IL81UMEC9OJI    2\n",
       "                 ..\n",
       "APB33P7IQEEWK     2\n",
       "AREQ43Q0T8ENT     2\n",
       "ASNMT3I94HZ44     2\n",
       "AT6ICEVY8RM8D     4\n",
       "ATJ92FYQX176Z     7\n",
       "Name: index, Length: 65, dtype: int64"
      ]
     },
     "execution_count": 601,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.concat(g for _, g in df.groupby('WorkerId') if len(g) > 1).groupby('WorkerId')['index'].count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "how many to drop?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 602,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "98"
      ]
     },
     "execution_count": 602,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum(pd.concat(g for _, g in df.groupby('WorkerId') if len(g) > 1).groupby('WorkerId')['index'].count()) - len(pd.concat(g for _, g in df.groupby('WorkerId') if len(g) > 1).groupby('WorkerId')['index'].count())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will keep the first one of any 'WorkerId' instance to keep samples unique"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 603,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[(~df.duplicated('WorkerId')) | (df['WorkerId'].isnull())]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 604,
   "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>aggressively_value</th>\n",
       "      <th>blueScore</th>\n",
       "      <th>computer_value</th>\n",
       "      <th>generously_value</th>\n",
       "      <th>redScore</th>\n",
       "      <th>steps</th>\n",
       "      <th>wisely_value</th>\n",
       "      <th>predictable_value</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>behavior</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>aggressiveBehavior</th>\n",
       "      <td>5.00</td>\n",
       "      <td>-22.38</td>\n",
       "      <td>5.96</td>\n",
       "      <td>3.30</td>\n",
       "      <td>-24.17</td>\n",
       "      <td>6.49</td>\n",
       "      <td>4.33</td>\n",
       "      <td>4.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>carefulBehavior</th>\n",
       "      <td>2.89</td>\n",
       "      <td>10.59</td>\n",
       "      <td>5.15</td>\n",
       "      <td>5.04</td>\n",
       "      <td>13.07</td>\n",
       "      <td>7.61</td>\n",
       "      <td>3.96</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>randomBehavior</th>\n",
       "      <td>3.51</td>\n",
       "      <td>-73.79</td>\n",
       "      <td>5.63</td>\n",
       "      <td>3.99</td>\n",
       "      <td>-73.12</td>\n",
       "      <td>12.06</td>\n",
       "      <td>3.74</td>\n",
       "      <td>3.51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>semiAggressiveBehavior</th>\n",
       "      <td>4.55</td>\n",
       "      <td>-61.81</td>\n",
       "      <td>5.74</td>\n",
       "      <td>3.09</td>\n",
       "      <td>-62.98</td>\n",
       "      <td>6.38</td>\n",
       "      <td>4.28</td>\n",
       "      <td>4.63</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>semiCarefulBehavior</th>\n",
       "      <td>3.87</td>\n",
       "      <td>-3.02</td>\n",
       "      <td>5.71</td>\n",
       "      <td>4.31</td>\n",
       "      <td>-1.27</td>\n",
       "      <td>11.80</td>\n",
       "      <td>4.91</td>\n",
       "      <td>4.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>valueIterationBasedBehavior_v2</th>\n",
       "      <td>4.89</td>\n",
       "      <td>-3.68</td>\n",
       "      <td>6.10</td>\n",
       "      <td>3.52</td>\n",
       "      <td>-5.60</td>\n",
       "      <td>8.61</td>\n",
       "      <td>4.77</td>\n",
       "      <td>4.86</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                aggressively_value  blueScore  computer_value  \\\n",
       "behavior                                                                        \n",
       "aggressiveBehavior                            5.00     -22.38            5.96   \n",
       "carefulBehavior                               2.89      10.59            5.15   \n",
       "randomBehavior                                3.51     -73.79            5.63   \n",
       "semiAggressiveBehavior                        4.55     -61.81            5.74   \n",
       "semiCarefulBehavior                           3.87      -3.02            5.71   \n",
       "valueIterationBasedBehavior_v2                4.89      -3.68            6.10   \n",
       "\n",
       "                                generously_value  redScore  steps  \\\n",
       "behavior                                                            \n",
       "aggressiveBehavior                          3.30    -24.17   6.49   \n",
       "carefulBehavior                             5.04     13.07   7.61   \n",
       "randomBehavior                              3.99    -73.12  12.06   \n",
       "semiAggressiveBehavior                      3.09    -62.98   6.38   \n",
       "semiCarefulBehavior                         4.31     -1.27  11.80   \n",
       "valueIterationBasedBehavior_v2              3.52     -5.60   8.61   \n",
       "\n",
       "                                wisely_value  predictable_value  \n",
       "behavior                                                         \n",
       "aggressiveBehavior                      4.33               4.92  \n",
       "carefulBehavior                         3.96                NaN  \n",
       "randomBehavior                          3.74               3.51  \n",
       "semiAggressiveBehavior                  4.28               4.63  \n",
       "semiCarefulBehavior                     4.91               4.32  \n",
       "valueIterationBasedBehavior_v2          4.77               4.86  "
      ]
     },
     "execution_count": 604,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('behavior').mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 605,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[df['behavior'] != 'valueIterationBasedBehavior_v2']"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Statistics of the red agent movements:"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create the statistics table from the log column: "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 606,
   "metadata": {},
   "outputs": [],
   "source": [
    "class OutOfBoard(Exception):\n",
    "    pass"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 607,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_pos_after_action(s, a):\n",
    "    if a == 'U' or a == \"up\":\n",
    "        if(s[0] == 'a'):\n",
    "            raise OutOfBoard(\"An error occurred, state: \"+ s + \" action: \" + a)\n",
    "        return \"a\" + s[1]\n",
    "    if a == 'D' or a == \"down\":\n",
    "        if(s[0] == 'b'):\n",
    "            raise OutOfBoard(\"An error occurred, state: \"+ s + \" action: \" + a)\n",
    "        return \"b\" + s[1]\n",
    "    if a == 'L' or a == \"left\":\n",
    "        if(s[0] == 'b' or s[1] == '1'):\n",
    "            raise OutOfBoard(\"An error occurred, state: \"+ s + \" action: \" + a)\n",
    "        return \"a\" + str(int(s[1])-1)\n",
    "    if a == 'R' or a == \"right\":\n",
    "        if(s[0] == 'b' or s[1] == '6'):\n",
    "            raise OutOfBoard(\"An error occurred, state: \"+ s + \" action: \" + a)\n",
    "        return \"a\" + str(int(s[1])+1)\n",
    "    if a == 'S' or a == \"stay\":\n",
    "        return s"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 608,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_next_state(c_s):\n",
    "    n_s = c_s.split()\n",
    "    blueAction = n_s[3]\n",
    "    blueState = n_s[2]\n",
    "    redAction = n_s[1]\n",
    "    redState = n_s[0]\n",
    "    \n",
    "    newPosBlue = get_pos_after_action(blueState, blueAction)\n",
    "    newPosRed = get_pos_after_action(redState, redAction)\n",
    "    \n",
    "    return newPosBlue + ' ' + newPosRed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 609,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "s_r = df['log'].apply(pd.Series).iloc[:,1:]\n",
    "s_r.columns = s_r.columns.astype(str)\n",
    "s_r.set_index( df['index'], inplace=True)\n",
    "log_red = flat_table.normalize(s_r)\n",
    "log_red = log_red.dropna(how='all')\n",
    "log_red.set_index(log_red['index'], inplace=True)\n",
    "log_red = log_red.drop('index', axis=1)\n",
    "\n",
    "log_len = int(len(log_red.columns)/2)+1\n",
    "for i in range(1,log_len):\n",
    "    log_red[str(i)] = log_red[str(i)+'.red'] + ' ' + log_red[str(i)+'.blue']\n",
    "    log_red = log_red.drop([str(i)+'.red', str(i)+'.blue'],axis=1)\n",
    "    for substr in ['from: ', ', action:']:\n",
    "        log_red[str(i)] = log_red[str(i)].apply(lambda x: x.replace(substr, '', 2) if type(x) != float else x)\n",
    "    \n",
    "log_len = len(log_red.columns)\n",
    "for i in range(1,log_len):\n",
    "    log_red[str(i) + '&' + str(i+1)] = log_red[str(i)] + '&' + log_red[str(i+1)]\n",
    "    log_red[str(i) + '&' + str(i+1)] = log_red[str(i) + '&' + str(i+1)].apply(lambda x: x.rsplit(' ', 2)[0] if type(x) != float else x)\n",
    "log_red = log_red.drop([str(i) for i in range(1,log_len+1)],axis=1)\n",
    "\n",
    "log_red_copy = log_red.copy(deep=True)\n",
    "\n",
    "log_red = pd.DataFrame(log_red.to_numpy().flatten(), columns=['state'])\n",
    "log_red = log_red.dropna()\n",
    "\n",
    "log_red['action'] = log_red['state'].apply(lambda x: x.rsplit('&', 1)[1].rsplit(' ', 1)[1])\n",
    "log_red['state'] = log_red['state'].apply(lambda x: x.rsplit('&', 1)[0])\n",
    "\n",
    "log_red['nxt'] = log_red['state'].apply(get_next_state)\n",
    "log_red['state'] = log_red['state'].str.split().apply(reversed).apply(' '.join)\n",
    "\n",
    "log_red['value']=1\n",
    "\n",
    "for a in ['right', 'stay', 'up', 'down', 'left']:\n",
    "    log_red['state'] = log_red['state'].apply(lambda x: x.replace(a+' ', '', 2))\n",
    "\n",
    "log_red['state'] = log_red['state'] + ' nxt: ' + log_red['nxt']\n",
    "log_red = log_red.drop('nxt',axis=1)\n",
    "\n",
    "log_red = log_red.groupby(['state', 'action']).count()\n",
    "log_red.reset_index(inplace=True)\n",
    "log_red = log_red.set_index('state')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 610,
   "metadata": {
    "scrolled": true
   },
   "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>state</th>\n",
       "      <th>right</th>\n",
       "      <th>stay</th>\n",
       "      <th>up</th>\n",
       "      <th>down</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>a1 a2 nxt: a1 a3</td>\n",
       "      <td>10.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>a1 a3 nxt: a1 a3</td>\n",
       "      <td>0.00</td>\n",
       "      <td>5.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>a1 a3 nxt: a1 a4</td>\n",
       "      <td>49.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>a1 a3 nxt: a1 b3</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>a1 a4 nxt: a1 a5</td>\n",
       "      <td>68.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>252</th>\n",
       "      <td>b6 b1 nxt: b6 a1</td>\n",
       "      <td>2.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>253</th>\n",
       "      <td>b6 b1 nxt: b6 b1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>254</th>\n",
       "      <td>b6 b4 nxt: a6 a4</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>255</th>\n",
       "      <td>b6 b5 nxt: b6 a5</td>\n",
       "      <td>23.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>4.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>256</th>\n",
       "      <td>b6 b5 nxt: b6 b5</td>\n",
       "      <td>0.00</td>\n",
       "      <td>2.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>257 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                state  right  stay   up  down\n",
       "0    a1 a2 nxt: a1 a3  10.00  0.00 0.00  0.00\n",
       "1    a1 a3 nxt: a1 a3   0.00  5.00 0.00  1.00\n",
       "2    a1 a3 nxt: a1 a4  49.00  0.00 0.00  0.00\n",
       "3    a1 a3 nxt: a1 b3   0.00  0.00 2.00  0.00\n",
       "4    a1 a4 nxt: a1 a5  68.00  0.00 0.00  0.00\n",
       "..                ...    ...   ...  ...   ...\n",
       "252  b6 b1 nxt: b6 a1   2.00  0.00 0.00  0.00\n",
       "253  b6 b1 nxt: b6 b1   0.00  2.00 0.00  0.00\n",
       "254  b6 b4 nxt: a6 a4   1.00  0.00 0.00  0.00\n",
       "255  b6 b5 nxt: b6 a5  23.00  3.00 0.00  4.00\n",
       "256  b6 b5 nxt: b6 b5   0.00  2.00 4.00  0.00\n",
       "\n",
       "[257 rows x 5 columns]"
      ]
     },
     "execution_count": 610,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result = pd.DataFrame(index=log_red.index.drop_duplicates(), columns=['right', 'stay', 'up', 'down'])\n",
    "# result.reset_index(level=0, inplace=True)\n",
    "for action in ['right', 'stay', 'up', 'down']:\n",
    "    result[action] = log_red[log_red['action'] == action]['value']\n",
    "result = result.fillna(0)\n",
    "result.reset_index(inplace=True)\n",
    "result.sort_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 611,
   "metadata": {},
   "outputs": [],
   "source": [
    "# laplace smoving:\n",
    "for i in range(1,7):\n",
    "    result.loc[result['state'].str.endswith('a'+str(i)), ['stay', 'right', 'down']] += 1\n",
    "    \n",
    "for i in range(1,7):\n",
    "    result.loc[result['state'].str.endswith('b'+str(i)), ['stay', 'up']] += 1\n",
    "\n",
    "result = result.set_index('state')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert it to Probability table:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 612,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "pd.options.display.float_format = \"{:,.2f}\".format\n",
    "result_probs = result.div(result.sum(axis=1), axis=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### model accuracy:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 613,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict(state):\n",
    "    score = result_probs.loc[state ,:]\n",
    "    dict_scores = dict(enumerate(score))\n",
    "    action = random.choices(list(dict_scores.keys()), weights=list(dict_scores.values()))[0]\n",
    "    return action\n",
    "\n",
    "def predicts(states):\n",
    "    y_pred = []\n",
    "    for s in states:\n",
    "        y_pred.append(predict(s))\n",
    "    return y_pred"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 614,
   "metadata": {},
   "outputs": [],
   "source": [
    "target = []\n",
    "for s in result_probs.index:\n",
    "    target.append( np.argmax(result_probs.loc[s ,:]))\n",
    "\n",
    "X_train,X_test,y_train,y_test = train_test_split(result_probs.index, target,test_size=0.20,random_state=43, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 615,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.6163846153846153"
      ]
     },
     "execution_count": 615,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "k_fold = 1000\n",
    "scores = []\n",
    "\n",
    "for _ in range(k_fold):\n",
    "\n",
    "    y_pred = predicts(X_test)\n",
    "    accuracy = metrics.accuracy_score(y_test, y_pred)\n",
    "    scores.append(accuracy)\n",
    "\n",
    "accuracy = np.mean(scores)\n",
    "# correlation = df.blueScore.corr(df.redScore)\n",
    "correlation = 0.74\n",
    "\n",
    "accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 616,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.74"
      ]
     },
     "execution_count": 616,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "correlation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 617,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.20054910769998768"
      ]
     },
     "execution_count": 617,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "1 - correlation/accuracy"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "e7370f93d1d0cde622a1f8e1c04877d8463912d04d973331ad4851f04de6915a"
  },
  "kernelspec": {
   "display_name": "Python 3.8.10 64-bit",
   "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": "3.8.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}