StateDetctionPart.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T11:53:46.402510Z",
     "start_time": "2020-07-31T11:53:42.855694Z"
    }
   },
   "source": [
    "# Hands State Detection"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T11:55:58.780672Z",
     "start_time": "2020-07-31T11:55:58.767670Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import glob\n",
    "from sklearn.decomposition import PCA\n",
    "import pandas.util.testing as tm\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading the data from disk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T11:53:46.417493Z",
     "start_time": "2020-07-31T11:53:46.404497Z"
    }
   },
   "outputs": [],
   "source": [
    "NUM_OF_RECORDS_PER_SAMPLE = 5\n",
    "def combine_rows(df):\n",
    "    cols = df.columns\n",
    "    df = df.groupby(df.index//NUM_OF_RECORDS_PER_SAMPLE)[cols].apply(lambda x: pd.DataFrame(x.values)).unstack().reset_index()\n",
    "    df.columns = df.columns.droplevel()\n",
    "    new_cols = [\"index\"]\n",
    "    for col in cols:\n",
    "        for i in range(NUM_OF_RECORDS_PER_SAMPLE):\n",
    "            new_cols.append(col+str(i)) \n",
    "        \n",
    "    df.columns = new_cols\n",
    "    not_important_features = [\"index\"]\n",
    "    features = [\"Time\",\"Frame ID\",\"state_type\",\"name\",\"Hand Type\",\"# hands\"]\n",
    "    for f in features:\n",
    "        for s in new_cols:\n",
    "            if f in s:\n",
    "                df[f] = df[s]\n",
    "                not_important_features.append(s)\n",
    "    df = df.drop(not_important_features,axis=1)\n",
    "    numeric_featers = [x for x in df.columns if x not in features]\n",
    "    df[numeric_featers] = df[numeric_featers].apply(pd.to_numeric)\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T11:53:46.430490Z",
     "start_time": "2020-07-31T11:53:46.421489Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_state_type(filename):\n",
    "    state_type = 3\n",
    "    if 'Spontan' in filename:\n",
    "        state_type = 1\n",
    "    elif 'Sync' in filename:\n",
    "        state_type = 2\n",
    "    return state_type\n",
    "\n",
    "def handle_right_hand_file(df):\n",
    "    alone_right = pd.read_csv(\"HandRight.csv\",index_col=None, header=0)\n",
    "    alone_right.rename(columns=lambda x: x.strip(), inplace=True)\n",
    "    alone_right[\"state_type\"] = 3\n",
    "    alone_right[\"name\"] = \"\"\n",
    "    #df = pd.concat([df,alone_right])\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T11:55:57.294424Z",
     "start_time": "2020-07-31T11:53:46.435489Z"
    }
   },
   "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>Position X0</th>\n",
       "      <th>Position X1</th>\n",
       "      <th>Position X2</th>\n",
       "      <th>Position X3</th>\n",
       "      <th>Position X4</th>\n",
       "      <th>Position Y0</th>\n",
       "      <th>Position Y1</th>\n",
       "      <th>Position Y2</th>\n",
       "      <th>Position Y3</th>\n",
       "      <th>Position Y4</th>\n",
       "      <th>...</th>\n",
       "      <th>Pinch Strength1</th>\n",
       "      <th>Pinch Strength2</th>\n",
       "      <th>Pinch Strength3</th>\n",
       "      <th>Pinch Strength4</th>\n",
       "      <th>Time</th>\n",
       "      <th>Frame ID</th>\n",
       "      <th>state_type</th>\n",
       "      <th>name</th>\n",
       "      <th>Hand Type</th>\n",
       "      <th># hands</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-110.42230</td>\n",
       "      <td>137.58480</td>\n",
       "      <td>-110.69090</td>\n",
       "      <td>138.65750</td>\n",
       "      <td>-111.44740</td>\n",
       "      <td>231.4766</td>\n",
       "      <td>206.9182</td>\n",
       "      <td>233.3635</td>\n",
       "      <td>210.6163</td>\n",
       "      <td>236.8768</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71.48033</td>\n",
       "      <td>98080</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>139.30040</td>\n",
       "      <td>-112.54570</td>\n",
       "      <td>139.50180</td>\n",
       "      <td>-113.14660</td>\n",
       "      <td>139.50440</td>\n",
       "      <td>214.3510</td>\n",
       "      <td>241.5983</td>\n",
       "      <td>218.3653</td>\n",
       "      <td>245.3283</td>\n",
       "      <td>222.8770</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71.51359</td>\n",
       "      <td>98084</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-113.91830</td>\n",
       "      <td>140.34820</td>\n",
       "      <td>-115.02970</td>\n",
       "      <td>142.05480</td>\n",
       "      <td>-115.33540</td>\n",
       "      <td>249.2464</td>\n",
       "      <td>227.4296</td>\n",
       "      <td>252.5950</td>\n",
       "      <td>233.2705</td>\n",
       "      <td>256.1827</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71.56316</td>\n",
       "      <td>98090</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>144.09590</td>\n",
       "      <td>-114.90850</td>\n",
       "      <td>145.90060</td>\n",
       "      <td>-115.70580</td>\n",
       "      <td>148.04590</td>\n",
       "      <td>239.4014</td>\n",
       "      <td>259.3652</td>\n",
       "      <td>243.7470</td>\n",
       "      <td>261.6903</td>\n",
       "      <td>247.7996</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71.59629</td>\n",
       "      <td>98094</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-116.08190</td>\n",
       "      <td>148.85690</td>\n",
       "      <td>-117.53250</td>\n",
       "      <td>150.12540</td>\n",
       "      <td>-118.87450</td>\n",
       "      <td>262.2935</td>\n",
       "      <td>249.4935</td>\n",
       "      <td>265.2766</td>\n",
       "      <td>252.2394</td>\n",
       "      <td>269.4034</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>71.64600</td>\n",
       "      <td>98099</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>left</td>\n",
       "      <td>2</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60027</th>\n",
       "      <td>91.31541</td>\n",
       "      <td>-74.79743</td>\n",
       "      <td>92.07137</td>\n",
       "      <td>-75.16435</td>\n",
       "      <td>92.72672</td>\n",
       "      <td>182.4873</td>\n",
       "      <td>153.2910</td>\n",
       "      <td>184.3723</td>\n",
       "      <td>154.5737</td>\n",
       "      <td>185.6738</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>855.99660</td>\n",
       "      <td>67290</td>\n",
       "      <td>2</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60028</th>\n",
       "      <td>-75.83507</td>\n",
       "      <td>93.94934</td>\n",
       "      <td>-76.47509</td>\n",
       "      <td>94.50477</td>\n",
       "      <td>-76.48546</td>\n",
       "      <td>159.3587</td>\n",
       "      <td>190.0024</td>\n",
       "      <td>163.2591</td>\n",
       "      <td>193.3147</td>\n",
       "      <td>165.9167</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>856.09640</td>\n",
       "      <td>67297</td>\n",
       "      <td>2</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60029</th>\n",
       "      <td>94.64355</td>\n",
       "      <td>-76.36160</td>\n",
       "      <td>95.38348</td>\n",
       "      <td>-76.51263</td>\n",
       "      <td>96.97344</td>\n",
       "      <td>196.8384</td>\n",
       "      <td>167.8417</td>\n",
       "      <td>199.8574</td>\n",
       "      <td>171.0094</td>\n",
       "      <td>203.6819</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>856.16330</td>\n",
       "      <td>67302</td>\n",
       "      <td>2</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60030</th>\n",
       "      <td>-77.62601</td>\n",
       "      <td>97.75611</td>\n",
       "      <td>-78.43058</td>\n",
       "      <td>98.93275</td>\n",
       "      <td>-79.84506</td>\n",
       "      <td>173.7391</td>\n",
       "      <td>205.9629</td>\n",
       "      <td>176.0170</td>\n",
       "      <td>208.3153</td>\n",
       "      <td>180.2403</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>856.26310</td>\n",
       "      <td>67309</td>\n",
       "      <td>2</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60031</th>\n",
       "      <td>101.91230</td>\n",
       "      <td>-81.57359</td>\n",
       "      <td>103.84750</td>\n",
       "      <td>-84.05124</td>\n",
       "      <td>106.11880</td>\n",
       "      <td>212.0264</td>\n",
       "      <td>183.5054</td>\n",
       "      <td>214.7484</td>\n",
       "      <td>187.8450</td>\n",
       "      <td>218.0190</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>856.32950</td>\n",
       "      <td>67314</td>\n",
       "      <td>2</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>60032 rows × 96 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Position X0  Position X1  Position X2  Position X3  Position X4  \\\n",
       "0       -110.42230    137.58480   -110.69090    138.65750   -111.44740   \n",
       "1        139.30040   -112.54570    139.50180   -113.14660    139.50440   \n",
       "2       -113.91830    140.34820   -115.02970    142.05480   -115.33540   \n",
       "3        144.09590   -114.90850    145.90060   -115.70580    148.04590   \n",
       "4       -116.08190    148.85690   -117.53250    150.12540   -118.87450   \n",
       "...            ...          ...          ...          ...          ...   \n",
       "60027     91.31541    -74.79743     92.07137    -75.16435     92.72672   \n",
       "60028    -75.83507     93.94934    -76.47509     94.50477    -76.48546   \n",
       "60029     94.64355    -76.36160     95.38348    -76.51263     96.97344   \n",
       "60030    -77.62601     97.75611    -78.43058     98.93275    -79.84506   \n",
       "60031    101.91230    -81.57359    103.84750    -84.05124    106.11880   \n",
       "\n",
       "       Position Y0  Position Y1  Position Y2  Position Y3  Position Y4  ...  \\\n",
       "0         231.4766     206.9182     233.3635     210.6163     236.8768  ...   \n",
       "1         214.3510     241.5983     218.3653     245.3283     222.8770  ...   \n",
       "2         249.2464     227.4296     252.5950     233.2705     256.1827  ...   \n",
       "3         239.4014     259.3652     243.7470     261.6903     247.7996  ...   \n",
       "4         262.2935     249.4935     265.2766     252.2394     269.4034  ...   \n",
       "...            ...          ...          ...          ...          ...  ...   \n",
       "60027     182.4873     153.2910     184.3723     154.5737     185.6738  ...   \n",
       "60028     159.3587     190.0024     163.2591     193.3147     165.9167  ...   \n",
       "60029     196.8384     167.8417     199.8574     171.0094     203.6819  ...   \n",
       "60030     173.7391     205.9629     176.0170     208.3153     180.2403  ...   \n",
       "60031     212.0264     183.5054     214.7484     187.8450     218.0190  ...   \n",
       "\n",
       "       Pinch Strength1  Pinch Strength2  Pinch Strength3  Pinch Strength4  \\\n",
       "0                  0.0              0.0              0.0              0.0   \n",
       "1                  0.0              0.0              0.0              0.0   \n",
       "2                  0.0              0.0              0.0              0.0   \n",
       "3                  0.0              0.0              0.0              0.0   \n",
       "4                  0.0              0.0              0.0              0.0   \n",
       "...                ...              ...              ...              ...   \n",
       "60027              0.0              0.0              0.0              0.0   \n",
       "60028              0.0              0.0              0.0              0.0   \n",
       "60029              0.0              0.0              0.0              0.0   \n",
       "60030              0.0              0.0              0.0              0.0   \n",
       "60031              0.0              0.0              0.0              0.0   \n",
       "\n",
       "            Time  Frame ID  state_type           name  Hand Type  # hands  \n",
       "0       71.48033     98080           1  Evyatar Cohen       left        2  \n",
       "1       71.51359     98084           1  Evyatar Cohen      right        2  \n",
       "2       71.56316     98090           1  Evyatar Cohen       left        2  \n",
       "3       71.59629     98094           1  Evyatar Cohen      right        2  \n",
       "4       71.64600     98099           1  Evyatar Cohen       left        2  \n",
       "...          ...       ...         ...            ...        ...      ...  \n",
       "60027  855.99660     67290           2       Yoel Raz      right        2  \n",
       "60028  856.09640     67297           2       Yoel Raz       left        2  \n",
       "60029  856.16330     67302           2       Yoel Raz      right        2  \n",
       "60030  856.26310     67309           2       Yoel Raz       left        2  \n",
       "60031  856.32950     67314           2       Yoel Raz      right        2  \n",
       "\n",
       "[60032 rows x 96 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "SKIP_MIN_NUM = 7\n",
    "def construct_datasets(path_type):\n",
    "    all_files = glob.glob(path_type)\n",
    "    li = []\n",
    "\n",
    "    for filename in all_files:\n",
    "        df = pd.read_csv(filename, index_col=None, header=0)\n",
    "        df.rename(columns=lambda x: x.strip(), inplace=True)\n",
    "        \n",
    "        df['state_type'] = get_state_type(filename)\n",
    "        df['name'] = filename.split(\"\\\\\")[1]\n",
    "        start_time = df[\"Time\"].min()\n",
    "        df = df[df[\"Time\"]>start_time+SKIP_MIN_NUM]\n",
    "        df = combine_rows(df)\n",
    "        df[\"Time\"] = df[\"Time\"].astype(\"float\")\n",
    "        df = df[df['state_type'].notna()]\n",
    "        df[\"state_type\"] = df[\"state_type\"].astype(\"int\")\n",
    "        li.append(df)\n",
    "        \n",
    "    df = pd.concat(li, axis=0, ignore_index=True)\n",
    "    df = handle_right_hand_file(df)\n",
    "    \n",
    "    return df\n",
    "df = construct_datasets(\"Training/*/*.csv\")\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Labels\n",
    "\n",
    "\n",
    "| Label | Description |\n",
    "| --- | --- |\n",
    "| 1 | Spontan |\n",
    "| 2 | Sync |\n",
    "| 3 | Alone |"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data cleaning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "333.60600    2\n",
       "737.55660    2\n",
       "293.18770    2\n",
       "347.99260    2\n",
       "347.21410    2\n",
       "            ..\n",
       "70.95866     1\n",
       "444.25310    1\n",
       "407.09910    1\n",
       "404.58470    1\n",
       "388.23350    1\n",
       "Name: Time, Length: 59753, dtype: int64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values('Time')\n",
    "df['Time'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 60032 entries, 0 to 60031\n",
      "Data columns (total 96 columns):\n",
      " #   Column           Non-Null Count  Dtype  \n",
      "---  ------           --------------  -----  \n",
      " 0   Position X0      60032 non-null  float64\n",
      " 1   Position X1      60032 non-null  float64\n",
      " 2   Position X2      60032 non-null  float64\n",
      " 3   Position X3      60032 non-null  float64\n",
      " 4   Position X4      60032 non-null  float64\n",
      " 5   Position Y0      60032 non-null  float64\n",
      " 6   Position Y1      60032 non-null  float64\n",
      " 7   Position Y2      60032 non-null  float64\n",
      " 8   Position Y3      60032 non-null  float64\n",
      " 9   Position Y4      60032 non-null  float64\n",
      " 10  Position Z0      60032 non-null  float64\n",
      " 11  Position Z1      60032 non-null  float64\n",
      " 12  Position Z2      60032 non-null  float64\n",
      " 13  Position Z3      60032 non-null  float64\n",
      " 14  Position Z4      60032 non-null  float64\n",
      " 15  Velocity X0      60032 non-null  float64\n",
      " 16  Velocity X1      60032 non-null  float64\n",
      " 17  Velocity X2      60032 non-null  float64\n",
      " 18  Velocity X3      60032 non-null  float64\n",
      " 19  Velocity X4      60032 non-null  float64\n",
      " 20  Velocity Y0      60032 non-null  float64\n",
      " 21  Velocity Y1      60032 non-null  float64\n",
      " 22  Velocity Y2      60032 non-null  float64\n",
      " 23  Velocity Y3      60032 non-null  float64\n",
      " 24  Velocity Y4      60032 non-null  float64\n",
      " 25  Velocity Z0      60032 non-null  float64\n",
      " 26  Velocity Z1      60032 non-null  float64\n",
      " 27  Velocity Z2      60032 non-null  float64\n",
      " 28  Velocity Z3      60032 non-null  float64\n",
      " 29  Velocity Z4      60032 non-null  float64\n",
      " 30  Pitch0           60032 non-null  float64\n",
      " 31  Pitch1           60032 non-null  float64\n",
      " 32  Pitch2           60032 non-null  float64\n",
      " 33  Pitch3           60032 non-null  float64\n",
      " 34  Pitch4           60032 non-null  float64\n",
      " 35  Roll0            60032 non-null  float64\n",
      " 36  Roll1            60032 non-null  float64\n",
      " 37  Roll2            60032 non-null  float64\n",
      " 38  Roll3            60032 non-null  float64\n",
      " 39  Roll4            60032 non-null  float64\n",
      " 40  Yaw0             60032 non-null  float64\n",
      " 41  Yaw1             60032 non-null  float64\n",
      " 42  Yaw2             60032 non-null  float64\n",
      " 43  Yaw3             60032 non-null  float64\n",
      " 44  Yaw4             60032 non-null  float64\n",
      " 45  Wrist Pos X0     60032 non-null  float64\n",
      " 46  Wrist Pos X1     60032 non-null  float64\n",
      " 47  Wrist Pos X2     60032 non-null  float64\n",
      " 48  Wrist Pos X3     60032 non-null  float64\n",
      " 49  Wrist Pos X4     60032 non-null  float64\n",
      " 50  Wrist Pos Y0     60032 non-null  float64\n",
      " 51  Wrist Pos Y1     60032 non-null  float64\n",
      " 52  Wrist Pos Y2     60032 non-null  float64\n",
      " 53  Wrist Pos Y3     60032 non-null  float64\n",
      " 54  Wrist Pos Y4     60032 non-null  float64\n",
      " 55  Wrist Pos Z0     60032 non-null  float64\n",
      " 56  Wrist Pos Z1     60032 non-null  float64\n",
      " 57  Wrist Pos Z2     60032 non-null  float64\n",
      " 58  Wrist Pos Z3     60032 non-null  float64\n",
      " 59  Wrist Pos Z4     60032 non-null  float64\n",
      " 60  Elbow pos X0     60032 non-null  float64\n",
      " 61  Elbow pos X1     60032 non-null  float64\n",
      " 62  Elbow pos X2     60032 non-null  float64\n",
      " 63  Elbow pos X3     60032 non-null  float64\n",
      " 64  Elbow pos X4     60032 non-null  float64\n",
      " 65  Elbow Pos Y0     60032 non-null  float64\n",
      " 66  Elbow Pos Y1     60032 non-null  float64\n",
      " 67  Elbow Pos Y2     60032 non-null  float64\n",
      " 68  Elbow Pos Y3     60032 non-null  float64\n",
      " 69  Elbow Pos Y4     60032 non-null  float64\n",
      " 70  Elbow Pos Z0     60032 non-null  float64\n",
      " 71  Elbow Pos Z1     60032 non-null  float64\n",
      " 72  Elbow Pos Z2     60032 non-null  float64\n",
      " 73  Elbow Pos Z3     60032 non-null  float64\n",
      " 74  Elbow Pos Z4     60032 non-null  float64\n",
      " 75  Grab Strenth0    60032 non-null  float64\n",
      " 76  Grab Strenth1    60032 non-null  float64\n",
      " 77  Grab Strenth2    60032 non-null  float64\n",
      " 78  Grab Strenth3    60032 non-null  float64\n",
      " 79  Grab Strenth4    60032 non-null  float64\n",
      " 80  Grab Angle0      60032 non-null  float64\n",
      " 81  Grab Angle1      60032 non-null  float64\n",
      " 82  Grab Angle2      60032 non-null  float64\n",
      " 83  Grab Angle3      60032 non-null  float64\n",
      " 84  Grab Angle4      60032 non-null  float64\n",
      " 85  Pinch Strength0  60032 non-null  float64\n",
      " 86  Pinch Strength1  60032 non-null  float64\n",
      " 87  Pinch Strength2  60032 non-null  float64\n",
      " 88  Pinch Strength3  60032 non-null  float64\n",
      " 89  Pinch Strength4  60032 non-null  float64\n",
      " 90  Time             60032 non-null  float64\n",
      " 91  Frame ID         60032 non-null  object \n",
      " 92  state_type       60032 non-null  int32  \n",
      " 93  name             60032 non-null  object \n",
      " 94  Hand Type        60032 non-null  object \n",
      " 95  # hands          60032 non-null  object \n",
      "dtypes: float64(91), int32(1), object(4)\n",
      "memory usage: 43.7+ MB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.dates as matdates\n",
    "\n",
    "def plot_position_over_time(position):\n",
    "    print(position,\"over time\")\n",
    "    for name in df[\"name\"].dropna().unique()[:1]:#you can remove the [] to print for all persons\n",
    "        print(\"plot position of\",name)\n",
    "        fig, ax = plt.subplots(nrows=3, ncols=3,figsize=(20,10))\n",
    "        for state_type in range(1,4):\n",
    "            x_pos_over_timer = df[(df[\"name\"]==name) & (df[\"state_type\"]==state_type)][[\"Time\",position,\"Hand Type\"]]\n",
    "            left_x_pos_over_timer = x_pos_over_timer[x_pos_over_timer[\"Hand Type\"]==\"left\"]\n",
    "            right_x_pos_over_timer = x_pos_over_timer[x_pos_over_timer[\"Hand Type\"]==\"right\"]\n",
    "            ax[state_type-1][0].plot(left_x_pos_over_timer[\"Time\"],left_x_pos_over_timer[position])\n",
    "            ax[state_type-1][0].plot(right_x_pos_over_timer[\"Time\"],right_x_pos_over_timer[position])\n",
    "            ax[state_type-1][1].plot(left_x_pos_over_timer[\"Time\"],left_x_pos_over_timer[position])\n",
    "            ax[state_type-1][2].plot(right_x_pos_over_timer[\"Time\"],right_x_pos_over_timer[position],\"m\")\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n"
     ]
    }
   ],
   "source": [
    "print(len(df[pd.isnull(df).any(axis=1)]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Correlation between features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 3600x2520 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "numeric_df = df.drop([\"Frame ID\",\"state_type\",\"name\",\"Hand Type\",\"# hands\"], axis=1) \n",
    "corr = numeric_df.corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x1080 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f = plt.figure(figsize=(25, 15))\n",
    "plt.matshow(corr, fignum=f.number)\n",
    "plt.xticks(range(numeric_df.shape[1]), numeric_df.columns, fontsize=10, rotation=80)\n",
    "plt.yticks(range(numeric_df.shape[1]), numeric_df.columns, fontsize=10)\n",
    "cb = plt.colorbar()\n",
    "cb.ax.tick_params(labelsize=15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:06:27.067673Z",
     "start_time": "2020-07-31T12:06:20.583509Z"
    }
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import xgboost as xgb\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.svm import SVC\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Embedding, LSTM, SpatialDropout1D\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from sklearn.metrics import classification_report,confusion_matrix,accuracy_score\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.ensemble import AdaBoostClassifier\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import cross_val_score,train_test_split,StratifiedKFold\n",
    "\n",
    "#like kfold but with equal distrabuted chunks\n",
    "import warnings\n",
    "#warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Shuffle the data**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:06:27.338636Z",
     "start_time": "2020-07-31T12:06:27.071638Z"
    }
   },
   "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>Position X0</th>\n",
       "      <th>Position X1</th>\n",
       "      <th>Position X2</th>\n",
       "      <th>Position X3</th>\n",
       "      <th>Position X4</th>\n",
       "      <th>Position Y0</th>\n",
       "      <th>Position Y1</th>\n",
       "      <th>Position Y2</th>\n",
       "      <th>Position Y3</th>\n",
       "      <th>Position Y4</th>\n",
       "      <th>...</th>\n",
       "      <th>Pinch Strength1</th>\n",
       "      <th>Pinch Strength2</th>\n",
       "      <th>Pinch Strength3</th>\n",
       "      <th>Pinch Strength4</th>\n",
       "      <th>Time</th>\n",
       "      <th>Frame ID</th>\n",
       "      <th>state_type</th>\n",
       "      <th>name</th>\n",
       "      <th>Hand Type</th>\n",
       "      <th># hands</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>153.69320</td>\n",
       "      <td>-99.63034</td>\n",
       "      <td>153.36750</td>\n",
       "      <td>-100.33610</td>\n",
       "      <td>152.26960</td>\n",
       "      <td>230.2679</td>\n",
       "      <td>141.9383</td>\n",
       "      <td>230.4591</td>\n",
       "      <td>143.8630</td>\n",
       "      <td>229.4772</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>181.0006</td>\n",
       "      <td>21460</td>\n",
       "      <td>1</td>\n",
       "      <td>Orya Kalmanovitz</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-11.65949</td>\n",
       "      <td>-11.18732</td>\n",
       "      <td>-11.59908</td>\n",
       "      <td>-14.37793</td>\n",
       "      <td>-16.39284</td>\n",
       "      <td>194.9421</td>\n",
       "      <td>199.6535</td>\n",
       "      <td>203.5603</td>\n",
       "      <td>208.9689</td>\n",
       "      <td>214.8690</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>564.7716</td>\n",
       "      <td>154919</td>\n",
       "      <td>3</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>left</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-53.65945</td>\n",
       "      <td>144.51660</td>\n",
       "      <td>-53.23605</td>\n",
       "      <td>144.55360</td>\n",
       "      <td>-53.01060</td>\n",
       "      <td>180.2209</td>\n",
       "      <td>222.0503</td>\n",
       "      <td>180.2553</td>\n",
       "      <td>221.7570</td>\n",
       "      <td>180.8817</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>766.0015</td>\n",
       "      <td>88681</td>\n",
       "      <td>2</td>\n",
       "      <td>Shelly Social_Nuero</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-102.14430</td>\n",
       "      <td>-106.66830</td>\n",
       "      <td>-109.63040</td>\n",
       "      <td>-110.26700</td>\n",
       "      <td>-109.98140</td>\n",
       "      <td>211.4936</td>\n",
       "      <td>209.1679</td>\n",
       "      <td>207.5930</td>\n",
       "      <td>202.7849</td>\n",
       "      <td>197.2902</td>\n",
       "      <td>...</td>\n",
       "      <td>0.064172</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>718.7756</td>\n",
       "      <td>56160</td>\n",
       "      <td>3</td>\n",
       "      <td>Yoel Raz</td>\n",
       "      <td>left</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>128.07230</td>\n",
       "      <td>-119.02290</td>\n",
       "      <td>126.49980</td>\n",
       "      <td>-119.91950</td>\n",
       "      <td>125.34700</td>\n",
       "      <td>206.8646</td>\n",
       "      <td>188.9501</td>\n",
       "      <td>206.3213</td>\n",
       "      <td>187.1628</td>\n",
       "      <td>206.0566</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>401.9668</td>\n",
       "      <td>136159</td>\n",
       "      <td>1</td>\n",
       "      <td>Evyatar Cohen</td>\n",
       "      <td>right</td>\n",
       "      <td>2</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",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60027</th>\n",
       "      <td>-204.76260</td>\n",
       "      <td>92.86594</td>\n",
       "      <td>-204.88930</td>\n",
       "      <td>92.14623</td>\n",
       "      <td>-205.53120</td>\n",
       "      <td>239.7147</td>\n",
       "      <td>234.2593</td>\n",
       "      <td>240.0523</td>\n",
       "      <td>233.2304</td>\n",
       "      <td>241.5378</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>460.6937</td>\n",
       "      <td>53172</td>\n",
       "      <td>2</td>\n",
       "      <td>Oriya Social_Nuero</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60028</th>\n",
       "      <td>141.82210</td>\n",
       "      <td>-116.61760</td>\n",
       "      <td>141.88200</td>\n",
       "      <td>-116.00760</td>\n",
       "      <td>142.41620</td>\n",
       "      <td>179.6343</td>\n",
       "      <td>176.5283</td>\n",
       "      <td>180.1647</td>\n",
       "      <td>176.0050</td>\n",
       "      <td>181.8188</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>571.8669</td>\n",
       "      <td>66601</td>\n",
       "      <td>1</td>\n",
       "      <td>Revital Marbel</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60029</th>\n",
       "      <td>135.03870</td>\n",
       "      <td>-166.97630</td>\n",
       "      <td>134.25270</td>\n",
       "      <td>-163.27340</td>\n",
       "      <td>133.11950</td>\n",
       "      <td>212.5394</td>\n",
       "      <td>208.9394</td>\n",
       "      <td>209.4848</td>\n",
       "      <td>202.8968</td>\n",
       "      <td>206.4600</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>666.7536</td>\n",
       "      <td>77856</td>\n",
       "      <td>2</td>\n",
       "      <td>Nofar Social_Nuero</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60030</th>\n",
       "      <td>224.92770</td>\n",
       "      <td>-154.13500</td>\n",
       "      <td>227.72890</td>\n",
       "      <td>-152.69950</td>\n",
       "      <td>230.12140</td>\n",
       "      <td>293.4737</td>\n",
       "      <td>174.3275</td>\n",
       "      <td>296.0541</td>\n",
       "      <td>174.2669</td>\n",
       "      <td>297.8920</td>\n",
       "      <td>...</td>\n",
       "      <td>0.789192</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.779378</td>\n",
       "      <td>1.0</td>\n",
       "      <td>516.4098</td>\n",
       "      <td>62087</td>\n",
       "      <td>2</td>\n",
       "      <td>Shahar Terner</td>\n",
       "      <td>right</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60031</th>\n",
       "      <td>-92.19800</td>\n",
       "      <td>70.16494</td>\n",
       "      <td>-90.26189</td>\n",
       "      <td>70.07745</td>\n",
       "      <td>-88.66868</td>\n",
       "      <td>274.8719</td>\n",
       "      <td>126.0023</td>\n",
       "      <td>273.6147</td>\n",
       "      <td>126.7081</td>\n",
       "      <td>271.7104</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>633.0641</td>\n",
       "      <td>73645</td>\n",
       "      <td>1</td>\n",
       "      <td>Revital Marbel</td>\n",
       "      <td>left</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>60032 rows × 96 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Position X0  Position X1  Position X2  Position X3  Position X4  \\\n",
       "0        153.69320    -99.63034    153.36750   -100.33610    152.26960   \n",
       "1        -11.65949    -11.18732    -11.59908    -14.37793    -16.39284   \n",
       "2        -53.65945    144.51660    -53.23605    144.55360    -53.01060   \n",
       "3       -102.14430   -106.66830   -109.63040   -110.26700   -109.98140   \n",
       "4        128.07230   -119.02290    126.49980   -119.91950    125.34700   \n",
       "...            ...          ...          ...          ...          ...   \n",
       "60027   -204.76260     92.86594   -204.88930     92.14623   -205.53120   \n",
       "60028    141.82210   -116.61760    141.88200   -116.00760    142.41620   \n",
       "60029    135.03870   -166.97630    134.25270   -163.27340    133.11950   \n",
       "60030    224.92770   -154.13500    227.72890   -152.69950    230.12140   \n",
       "60031    -92.19800     70.16494    -90.26189     70.07745    -88.66868   \n",
       "\n",
       "       Position Y0  Position Y1  Position Y2  Position Y3  Position Y4  ...  \\\n",
       "0         230.2679     141.9383     230.4591     143.8630     229.4772  ...   \n",
       "1         194.9421     199.6535     203.5603     208.9689     214.8690  ...   \n",
       "2         180.2209     222.0503     180.2553     221.7570     180.8817  ...   \n",
       "3         211.4936     209.1679     207.5930     202.7849     197.2902  ...   \n",
       "4         206.8646     188.9501     206.3213     187.1628     206.0566  ...   \n",
       "...            ...          ...          ...          ...          ...  ...   \n",
       "60027     239.7147     234.2593     240.0523     233.2304     241.5378  ...   \n",
       "60028     179.6343     176.5283     180.1647     176.0050     181.8188  ...   \n",
       "60029     212.5394     208.9394     209.4848     202.8968     206.4600  ...   \n",
       "60030     293.4737     174.3275     296.0541     174.2669     297.8920  ...   \n",
       "60031     274.8719     126.0023     273.6147     126.7081     271.7104  ...   \n",
       "\n",
       "       Pinch Strength1  Pinch Strength2  Pinch Strength3  Pinch Strength4  \\\n",
       "0             0.000000              0.0         0.000000              0.0   \n",
       "1             0.000000              0.0         0.000000              0.0   \n",
       "2             0.000000              0.0         0.000000              0.0   \n",
       "3             0.064172              0.0         0.000000              0.0   \n",
       "4             0.000000              1.0         0.000000              1.0   \n",
       "...                ...              ...              ...              ...   \n",
       "60027         0.000000              0.0         0.000000              0.0   \n",
       "60028         0.000000              0.0         0.000000              0.0   \n",
       "60029         0.000000              0.0         0.000000              0.0   \n",
       "60030         0.789192              1.0         0.779378              1.0   \n",
       "60031         0.000000              0.0         0.000000              0.0   \n",
       "\n",
       "           Time  Frame ID  state_type                 name  Hand Type  # hands  \n",
       "0      181.0006     21460           1     Orya Kalmanovitz      right        2  \n",
       "1      564.7716    154919           3        Evyatar Cohen       left        1  \n",
       "2      766.0015     88681           2  Shelly Social_Nuero       left        2  \n",
       "3      718.7756     56160           3             Yoel Raz       left        1  \n",
       "4      401.9668    136159           1        Evyatar Cohen      right        2  \n",
       "...         ...       ...         ...                  ...        ...      ...  \n",
       "60027  460.6937     53172           2   Oriya Social_Nuero       left        2  \n",
       "60028  571.8669     66601           1       Revital Marbel      right        2  \n",
       "60029  666.7536     77856           2   Nofar Social_Nuero      right        2  \n",
       "60030  516.4098     62087           2        Shahar Terner      right        2  \n",
       "60031  633.0641     73645           1       Revital Marbel       left        2  \n",
       "\n",
       "[60032 rows x 96 columns]"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.sample(frac=1).reset_index(drop=True)\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For tuning the hyper parameters,let's **split** the train data(we don't want to touch the validation till end)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:06:27.487682Z",
     "start_time": "2020-07-31T12:06:27.344636Z"
    }
   },
   "outputs": [],
   "source": [
    "y = df[\"state_type\"]\n",
    "X = df.drop([\"Frame ID\",\"state_type\",\"name\",\"Hand Type\",\"# hands\"], axis=1)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Features Balance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:06:27.500637Z",
     "start_time": "2020-07-31T12:06:27.489635Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1    24385\n",
      "2    24195\n",
      "3    11452\n",
      "Name: state_type, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(y.value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Features importance using PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PC0</td>\n",
       "      <td>Elbow pos X2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>PC1</td>\n",
       "      <td>Velocity Y2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PC2</td>\n",
       "      <td>Elbow pos X1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>PC3</td>\n",
       "      <td>Velocity Z2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PC4</td>\n",
       "      <td>Velocity Y1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>PC5</td>\n",
       "      <td>Time</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>PC6</td>\n",
       "      <td>Time</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>PC7</td>\n",
       "      <td>Velocity Z1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>PC8</td>\n",
       "      <td>Elbow Pos Y2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>PC9</td>\n",
       "      <td>Velocity X3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>PC10</td>\n",
       "      <td>Elbow Pos Z2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>PC11</td>\n",
       "      <td>Elbow Pos Y2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>PC12</td>\n",
       "      <td>Elbow Pos Y1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>PC13</td>\n",
       "      <td>Velocity Y4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       0             1\n",
       "0    PC0  Elbow pos X2\n",
       "1    PC1   Velocity Y2\n",
       "2    PC2  Elbow pos X1\n",
       "3    PC3   Velocity Z2\n",
       "4    PC4   Velocity Y1\n",
       "5    PC5          Time\n",
       "6    PC6          Time\n",
       "7    PC7   Velocity Z1\n",
       "8    PC8  Elbow Pos Y2\n",
       "9    PC9   Velocity X3\n",
       "10  PC10  Elbow Pos Z2\n",
       "11  PC11  Elbow Pos Y2\n",
       "12  PC12  Elbow Pos Y1\n",
       "13  PC13   Velocity Y4"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pca = PCA(n_components=0.95)\n",
    "pca.fit(X_train)\n",
    "X_pc = pca.transform(X_train)\n",
    "n_pcs= pca.components_.shape[0]\n",
    "most_important = [np.abs(pca.components_[i]).argmax() for i in range(n_pcs)]\n",
    "initial_feature_names = list(X_train.columns)\n",
    "most_important_names = [initial_feature_names[most_important[i]] for i in range(n_pcs)]\n",
    "dic = {'PC{}'.format(i): most_important_names[i] for i in range(n_pcs)}\n",
    "pd.DataFrame(dic.items())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Simple ML models"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### k nearest neighbors(knn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       0.97      0.95      0.96      8041\n",
      "           2       0.95      0.97      0.96      8016\n",
      "           3       0.99      1.00      1.00      3754\n",
      "\n",
      "    accuracy                           0.97     19811\n",
      "   macro avg       0.97      0.97      0.97     19811\n",
      "weighted avg       0.97      0.97      0.97     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "knn = KNeighborsClassifier(n_neighbors=5,p=1,weights=\"distance\")\n",
    "knn.fit(X_train,y_train)\n",
    "y_pred = knn.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(confusion_matrix(y_test, y_pred), cmap=\"YlGnBu\",annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "dst = DecisionTreeClassifier(max_depth=12,random_state=0)\n",
    "dst.fit(X_train,y_train)\n",
    "y_pred = dst.predict(X_test)\n",
    "print(classification_report(y_pred,y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Boosting"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Extreme Gradient Boosting(xgboost)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_xgboost_model(x,y,verbose_eval=150):\n",
    "    param_list = [(\"eta\", 0.08), (\"max_depth\", 6), (\"subsample\", 0.8), \n",
    "                  (\"colsample_bytree\", 0.8), (\"objective\", \"multi:softmax\"), (\"eval_metric\", \"merror\"),\n",
    "                  (\"alpha\", 7), (\"lambda\", 2), (\"num_class\", 4)]\n",
    "    \n",
    "    model = xgb.XGBClassifier(param_list,random_state=93)\n",
    "    model.fit(x,y)\n",
    "    return model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       1.00      0.99      1.00      8051\n",
      "           2       1.00      1.00      1.00      8042\n",
      "           3       1.00      1.00      1.00      3718\n",
      "\n",
      "    accuracy                           1.00     19811\n",
      "   macro avg       1.00      1.00      1.00     19811\n",
      "weighted avg       1.00      1.00      1.00     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "xgb_model = get_xgboost_model(X_train,y_train)\n",
    "y_pred = xgb_model.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "feature_important = xgb_model.get_booster().get_score(importance_type='weight')\n",
    "keys = list(feature_important.keys())\n",
    "values = list(feature_important.values())\n",
    "\n",
    "data = pd.DataFrame(data=values, index=keys, columns=[\"score\"]).sort_values(by = \"score\", ascending=False)\n",
    "data.plot(kind='barh')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Adaboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:57:35.891434Z",
     "start_time": "2020-07-31T12:19:38.489694Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       0.98      0.98      0.98      8048\n",
      "           2       0.99      0.98      0.98      7974\n",
      "           3       0.99      1.00      0.99      3789\n",
      "\n",
      "    accuracy                           0.98     19811\n",
      "   macro avg       0.99      0.99      0.99     19811\n",
      "weighted avg       0.98      0.98      0.98     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "base_estimetor = DecisionTreeClassifier(max_depth=12,random_state=0)\n",
    "adboost = AdaBoostClassifier(base_estimetor,learning_rate=0.08,n_estimators=500, random_state=93)\n",
    "adboost.fit(X_train,y_train)\n",
    "y_pred = adboost.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T12:57:36.161437Z",
     "start_time": "2020-07-31T12:57:35.896434Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1458d661f88>"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWcAAAD4CAYAAAAw/yevAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAeA0lEQVR4nO3deXxU5fXH8c+ZbIDIEmrYVUBwQQEVFMUqiyyugK2KrZWqNfZVkSpVAbX6qxQVW0Vs1TZUqtYFlVpBQS2CVluVTUFBqiCLhi3KIqBASDi/P+YSBwnJpCSZm8v37et5Zea5z7333BEOT85dxtwdEREJl1iqAxARkb0pOYuIhJCSs4hICCk5i4iEkJKziEgIpVf1DmofeokuB6liX6/8dapDiLyYVflfFQGgne3vFiqSc7Z99vR+76+qaOYsIhJCmg6ISKSYRWPOqeQsIpESlRJUNI5CRCSgmbOISAiZhfYcX4UoOYtIxGjmLCISOipriIiEUFSSczSOQkQkELP0pFtZzOxIM5uf0Dab2XVm9n9mtiqh/+yEdUaa2VIz+9jM+ib09wv6lprZiGSOQzNnEYmUypo5u/vHQKf4Ni0NWAX8A7gcGOvuv99zv3YMMAhoDzQDXjOzdsHiB4HeQD4wx8ymuPtHZe1fyVlEIqWKyhq9gE/dfWUZV4P0Bya6+w5guZktBU4Kli1192Xx+GxiMLbM5KyyhohEilXkP7NcM5ub0HL3sdlBwNMJ74eY2QdmNsHMGgZ9zYHPE8bkB3376i+TkrOIRIpZLOnm7nnu3jmh5e29PcsEzgeeC7oeBtoQL3msAe7dPbSUcLyM/jKprCEikRKLVXpaOwt4z93XAez+CWBm44GXgrf5QMuE9VoAq4PX++rfJ82cRSRiYhVoSbmEhJKGmTVNWDYQWBi8ngIMMrMsM2sFtAVmA3OAtmbWKpiFDwrGlkkzZxGJlMo8IWhmdYhfZXF1Qvc9ZtaJeGlixe5l7r7IzJ4lfqKvCLjG3YuD7QwBXgXSgAnuvqi8fSs5i0ikVGZydvdvgEbf6ftJGeNHA6NL6Z8GTKvIvpWcRSRSLCLVWiVnEYmUqNy+reQsIpESi6WlOoRKoeQsIpGisoaISAiprCEiEkJKziIiIaSyhohICFnl376dEtE4ChGRgL7gVUQkhFTWEBEJIZ0QFBEJI5U1RERCKBoTZyVnEYmYWDSy8wGRnNu2bsrfHhxa8r7VoTmMum8S/3p7EX+480oOOqgWK/O/4PKhD7Jl6zZ6fv84Ro0YRGZGOoU7i7h59FP86+3441cnPz6CJjkNSE9P4z+z/8t1t05g165yv3HmgHPLzX/gjTfmkt2oPi+++AAAmzZtYdiwe1m1qoDmzXMYO/YG6tevW7LOhx8uYdDFI7jvvl/Rt9+pqQq9Rlqz5gtuumksX365kVjMuOiifgwefD4vv/xv/vjHp/j003yee+5ejjuubapDrXrRyM1ROYyyLVm2hq5njaTrWSM59Zyb+WZbIVNemcPD9+Ry690T6dJnOFNemcv1V58LwPoNW/jhFb+nS5/hXHX9w0y4/xcl27r0F+M4ud8ITjzzRg7JPpgfnNM1VYcVagMG9iRv/G179I0f/zyndD2OV199iFO6Hsf48c+XLCsuLube3z9Ot9M6VXeokZCWlsaIEVfw8ssP88wzv+epp6aydOlntGt3GH/4w8106dI+1SFWGzdLuoXZAZGcE/XodizLP1vHZ6u+pG3rpvx71mIAZr71AQPOjn+L+YJFK1izbiMAH32ST1ZWBpmZ8V8ytmzdBkB6ehoZmel4+d/TeEDq0qU9DeofvEffzBmz6T+gBwD9B/RgxmuzSpY98cQ0evc5hUbZ9as1zqjIycmmffsjAKhbtw6tW7dk3br1tGnTktatW6Q4umpmFWghVm5yNrOjzGy4mT1gZuOC10dXR3BV4cLzT+XZyW8D8NHH+Zzb+0QALjinKy2aNtpr/MCzT2LBohUUFhaV9E352wg+e/9PbN26neenztprHSnd+vWbyMnJBuLJZMOGrwBYt249r01/l0GD+qYyvMjIz1/H4sWf0rHjkakOJTVilnwLsTKTs5kNByYS/zdm9xcVGvC0mY2o+vAqV0ZGGuf0PrEkoV5945+5enAf/jN1NHXr1qZwZ9Ee449u14LfjvwRQ0b+ZY/+839yN606/4KszHS6dzu22uKPqrvufIRf3XAZaWnReA5vKn399TaGDr2Lm2++irp166Q6nNQwS76FWHknBK8E2rv7zsROM7sPWATcXdpKZpYL5AKkN+xMet0jKiHU/de3eyfmL1xOwZfxGdsnn67mvEvvAuCIVk04q+e39c7mTbJ5Jm8YP7v+IZavLNhrWzt27OSl197jvN4nMvOtD6vnAGq4Ro0aUFCwgZycbAoKNpAdlDAWLvyUXw27F4ifNHzzzXmkpadx5pknpzLcGmfnziKGDr2L887rTp8+B/AJ1bRwJ91klVfW2AU0K6W/abCsVO6e5+6d3b1zWBIzwEX9vy1pABzSqB4Qvxd/xNCBjH9iBgD169Xh+Udv4rYxE3ln7icl4w+qk0WTnAYApKXF6NejEx9/uroaj6Bm69mzC5NfeB2AyS+8Ts9e8Rr/azP+zIyZecyYmUefPqdw221XKzFXkLtzyy0P0Lp1Sy6/fECqw0mtA2TmfB0ww8yWAJ8HfYcCRwBDqjKwyla7ViY9v3/cHiWKi/qfytWX9QFg8iuzefzZNwD4+eC+tDm8MSOGDmTE0IEAnHfpXZgZkx65gczMDNLSYvzrP4sY/8Rr1X4sNcGvht3L7DmL2LRxM93P+BlDrh3Ez666gGHX/55Jf59Bs6bfY+z9N6Y6zMiYN+8jJk9+nXbtDqd///hlo8OGXUZh4U5GjfozGzZ8xdVX38HRR7fikUfuSHG0VSzcOTdp5l721QYWv1H9JKA58cPOB+a4e3EyO6h96CW6nKGKfb3y16kOIfJidkDcEhAC7fY7tbbtNyHpnLPklStCm8rL/RPn7ruAd6shFhGR/RfadFsxmg6ISKR4WjRu31ByFpFo0cxZRCSEQn4VRrKUnEUkWkJ+51+ylJxFJFqikZuVnEUkYlTWEBEJoYjcvq3kLCLREpGZczQuCBQR2a0Sn+dsZg3MbJKZ/dfMFpvZKWaWbWbTzWxJ8LNhMNaCRysvNbMPzOyEhO0MDsYvMbPByRyGkrOIRIrHLOmWhHHAK+5+FNARWAyMAGa4e1tgRvAe4CygbdBygYcBzCwbuB04mfijMG7fndDLouQsItFSSU+lM7N6wOnAIwDuXujum4D+wGPBsMeA3Y8B7A887nHvAg3MrCnQF5ju7hvcfSMwHehX3mEoOYtItFSgrGFmuWY2N6HlJmypNfAF8Fcze9/M/mJmBwGN3X0NQPAzJxjfnG+f3gnxh8Q1L6O/TDohKCLRUoFna7h7HpC3j8XpwAnAte4+y8zG8W0JozSlTcW9jP4yaeYsItFSeScE84F8d9/9RaGTiCfrdUG5guBnQcL4lgnrtwBWl9FfJiVnEYmWSvqCV3dfC3xuZru/KbcX8BEwBdh9xcVgYHLwegpwWXDVRlfgq6Ds8SrQx8waBicC+wR9ZVJZQ0SipXKfrXEt8KSZZQLLgMuJT2qfNbMrgc+AC4Ox04CzgaXAN8FY3H2DmY0i/gXZAHe4+4bydqzkLCKR4pWYm919PtC5lEW9ShnrwDX72M4EYEJF9q3kLCLRoofti4iEkB4ZKiISQtGYOCs5i0jEROTBR0rOIhItKmuIiISPa+YsIhJC6UrOIiLho5mziEgIqeYsIhJC0cjNSs4iEi1JfsNJ6Ck5i0i0KDmLiIRQmpJzUrauvKWqd3HAq9dqTKpDiLytK/TnuMbQ1RoiIiGksoaISAgpOYuIhI9u3xYRCSOdEBQRCSGVNUREQkjJWUQkhKKRm5WcRSRadPu2iEgY6WoNEZEQ0tUaIiLhE9O3b4uIhE9EqhpKziISLUrOIiIhZBHJzkrOIhIpqjmLiISQKTmLiIRPRKoaSs4iEi0RuUGQiPwCICISZ5Z8S257lmZm75vZS8H7R81suZnND1qnoN/M7AEzW2pmH5jZCQnbGGxmS4I2OJn9auYsIpFSBWWNXwKLgXoJfTe6+6TvjDsLaBu0k4GHgZPNLBu4HegMODDPzKa4+8aydqqZs4hESizNkm7lMbMWwDnAX5LYdX/gcY97F2hgZk2BvsB0d98QJOTpQL9yjyOJHYqI1BgVKWuYWa6ZzU1oud/Z3P3ATcCu7/SPDkoXY80sK+hrDnyeMCY/6NtXf5mUnEUkUiqSnN09z907J7S8b7dj5wIF7j7vO7sYCRwFdAGygeG7VyklHC+jv0xKziISKZV4QrAbcL6ZrQAmAj3N7Al3XxOULnYAfwVOCsbnAy0T1m8BrC6jv0xKziISKTFLvpXF3Ue6ewt3PxwYBMx090uDOjIWv098ALAwWGUKcFlw1UZX4Ct3XwO8CvQxs4Zm1hDoE/SVSVdriEikVMNNKE+a2SHEyxXzgZ8H/dOAs4GlwDfA5QDuvsHMRgFzgnF3uPuG8nai5CwikZLMVRgV5e5vAG8Er3vuY4wD1+xj2QRgQkX2qeQsIpGi27dFREJIyVlEJISUnEVEQigqDz5SchaRSImlpTqCynHAJedbbn6Qf70xl+xG9Zny4v0l/U/8bRpPPfkyaekxzjjjRG648TI++GAJt9/2p/gAd64ZcjFn9j45RZGHW9vWTXjsj78oeX94yxx+O/Z53nr3v4wbPZhaWRkUFe3i+l8/zrwFywD43e0/pk+PjmzbVsjVN4xnwaKVtGzeiKf+NJS0NCMjPZ0/PTadR558PVWHVWM9+ugLPPfcPzEz2rU7nLvu+iVZWZmpDqtaqKxRQw0c2J0f//gsRox4oKRv1rsfMnPmbF6Ych+ZmRmsX/8VAG3bHspzk+4hPT2NLwo2MnDAMLr36Ex6ekT+aa5ES5at5dSzbwMgFjOWzLqfF1+dxx/vuoK7xk1m+hsf0Kd7B3478iLOGnQ3fbp3oE2rJnTsfhNdjm/D/aMH02PAHawt2ESvH4yisLCIg+pkMfufo5k6/X3WFmxK8RHWHOvWrefxx19k2rSHqFUri1/+8m6mTn2TCy44M9WhVYuofIfgAXeHYOcu7alfv+4efRMnvsrPrhpIZmYGAI0a1Qegdu2skkS8o7AwMv/Tq1r3bu1ZtvILPl+1HsepV7cWAPXr1WHNuniSPbfPCTz9/H8AmPP+p9Q/uA6ND6nPzp3FFBYWAZCVmU4sKt85VM2Ki3exfXshRUXFbN++g5yc7FSHVG0q+3nOqfI/z5zN7HJ3/2tlBpMqK1asYd7cxYy7/2myMjO4cfhgjjvuCAAWLPiEW295kNWrv2TMmKGaNSfhh+edzKQp7wIw/DdP8sLjNzL65kHEYjF6/WAUAE0bNyR/9fqSdVav3UCzJg1Z98VXNG+azd8nDKP14TnceuczmjVXUOPGjbjiioH06HEFWVmZdOt2PKeddkL5K0ZE2JNusvZnWvKbfS1IfAzf+Lzn9mMX1aO4uJjNm79m4jN3ccNNlzHsunuJ3+wDHTu248WXxvHsc2MYn/c8O3YUpjjacMvISOOcM4/nH9NmA/CzS3syYtRTHHXqMEaMeoqHxlwJlP4XKPjIWbVmA13PupUOZ9zEj35wGjnfq7f3YNmnr77ayowZs5gx4y+89dZjbNu2ncmTD5y6fVRmzmUm5+B5paW1D4HG+1ov8TF8V+VeWOlBV7YmjRvRu/fJmBkdOrQlFjM2bty8x5g2bVpQu3YWSz75LEVR1gx9undg/sKVFHwZ//x+9IPTmPzKXACenzqbEzu2BmD12o20aNaoZL1mTbJZs27PL4ZYW7CJxUtWcWqXdtUUfTS8/fZ8WrRoTHZ2fTIy0unT51Tef39xqsOqNumx5FuYlRdeY+Ay4LxS2voy1qtRep55ErNmfQjAiuWr2bmziIYN65Gfv46iomIAVq0qYPny1TRvkZPKUEPvwvO78tyL75a8X1uwie93PQqA7qcew6cr1gEwdfr7XHJBNwC6HN+GzVu2se6Lr2jWpCG1suK1/wb16tD1xLYsWba2mo+iZmvW7BAWLPgv27Ztx915550FtGnTsvwVIyJmnnQLs/Jqzi8Bdd19/ncXmNkbVRJRFbth2H3MnrOITRu30OOMqxhy7cVccEFPbr3lIc4/7zoyMtK58+5rMTPem7eY8eP/QXp6OrGY8evbr6JhQ/2KvS+1a2XS47RjGXrzoyV9Q0ZM4J7bLyU9Pcb2HTu5dmT8NMWrry+gb48OfPCv37Ft2w5+fmP8W4COPKIZd91yCY5jGA+Mf5lFH+en4nBqrI4dj6Rv324MHHgd6elpHH10ay6+uNxvRYqMqNyEYrtrq1Wl2BeG+5+nCKjf6nepDiHytq64JdUhHCDa7XdqPeef/04650ztc1poU/kBd52ziERb2MsVyVJyFpFIiUpZQ8lZRCIlXclZRCR8TGUNEZHwUVlDRCSEQn5vSdKUnEUkUnS1hohICOmEoIhICKnmLCISQipriIiEkGbOIiIhpKs1RERCSGUNEZEQCvtD9JOl5CwikRKR3KzkLCLRorKGiEgI6WoNEZEQikpZIyrHISICxGfOybaymFktM5ttZgvMbJGZ/Sbob2Vms8xsiZk9Y2aZQX9W8H5psPzwhG2NDPo/NrO+SR3H//oBiIiEUVrMk27l2AH0dPeOQCegn5l1BcYAY929LbARuDIYfyWw0d2PAMYG4zCzY4BBQHugH/CQmaWVt3MlZxGJlFgFWlk8bmvwNiNoDvQEJgX9jwEDgtf9g/cEy3uZmQX9E919h7svB5YCJyVzHCIikREzT7qVx8zSzGw+UABMBz4FNrl7UTAkH2gevG4OfA4QLP8KaJTYX8o6+z6OpI5WRKSGqEjN2cxyzWxuQstN3Ja7F7t7J6AF8dnu0aXscneWL62K7WX0l0lXa4hIpFTkUjp3zwPykhi3yczeALoCDcwsPZgdtwBWB8PygZZAvpmlA/WBDQn9uyWus+/jSP4wRETCL8M86VYWMzvEzBoEr2sDZwKLgdeBHwbDBgOTg9dTgvcEy2e6uwf9g4KrOVoBbYHZ5R2HZs4iEimVeBNKU+Cx4MqKGPCsu79kZh8BE83st8D7wCPB+EeAv5nZUuIz5kEA7r7IzJ4FPgKKgGvcvbi8nSs5i0ikVFZydvcPgONL6V9GKVdbuPt24MJ9bGs0MLoi+1dyFpFISdPt2yIi4aNna4iIhJCeSiciEkIZmjknJy3+TBCpQltX3JLqECKv/2vrUh3CAWHyme32exsqa4iIhJDKGiIiIaSrNUREQkhlDRGRENK3b4uIhFCaas4iIuETkYmzkrOIRItqziIiIaTkLCISQqo5i4iEkK7WEBEJIZU1RERCSHcIioiEkJ6tISISQhEpOSs5i0i0qOYsIhJCGTGVNUREQkczZxGREFJyFhEJIZ0QFBEJIdPMWUQkfFTWEBEJIZU1RERCyHSHoIhI+ESkqqHkLCLRohOCIiIhFJHcrOQsItGiR4aKiIRQVMoaUbnqREQEiJc1km3lbstsgpkVmNnChL7/M7NVZjY/aGcnLBtpZkvN7GMz65vQ3y/oW2pmI5I5DiVnEYmUykzOwKNAv1L6x7p7p6BNAzCzY4BBQPtgnYfMLM3M0oAHgbOAY4BLgrFlUllDRCKlMu8QdPc3zezwJIf3Bya6+w5guZktBU4Kli1192UAZjYxGPtRWRs74JPzyJHjeOONOTRqVJ+XXnoQgE2btnD99fewatU6mjdvzP33D6d+/bopjjQali3L5/rr7yl5//nnaxk69Mf89Kf9UxhVzbBr504+GzsGLyrCi3dx8PEncsi5/Vl53xh2bd8OQPHWzdQ6rBUtrh7C+umvsHnOLAB8VzGFa9fQdsxYdhUWsuaxRyja/BVYjAannU52jzNTeWiVqiK52cxygdyErjx3z0ti1SFmdhkwF/iVu28EmgPvJozJD/oAPv9O/8nl7eCAT84XXNCLSy89h+HDx5b05eVN4pRTOpCbeyF5ec+RlzeJG2/8aeqCjJDWrVswefIDABQXF3P66T+ld+9TUhxVzWDp6Rw69AZitWrhxUWsvHcMddsfy2HDhpeMyR//EAd36ARAo979aNQ7/hv5lg/ns3Hma6QdVBcv2kTOBRdR69DDKN6+nRVjRnHQUceQ1bRZSo6rslXkOwSDRJxMMk70MDAK8ODnvcAVlP7vglN6+bjcIA/4mnOXLsdSv/7Be/TNmDGLAQN6ATBgQC9ee+3d0laV/fTOOwto2bIpzZvnpDqUGsHMiNWqBYAXF+O7iknMB8Xbt/PNx/+lbofj91p3y9zZ1Osc/w07vX4Dah16GABptWqR1bgpRZs2Vv0BVBOz5Nv/wt3XuXuxu+8CxvNt6SIfaJkwtAWwuoz+MpWbnM3sKDPrZWZ1v9NfWpE8Etav30ROTjYAOTnZbNiwKcURRdPUqW9x7rmnpzqMGsV37WL5nb9hyfBhHHTUMdRu1bpk2dYF73HQkUeTVrv2HuvsKtzB1o8WcnCnE/baXuH6L9me/xm1Dm+917KaKlaB9r8ws6YJbwcCu6/kmAIMMrMsM2sFtAVmA3OAtmbWyswyiZ80nJLMcZQVxFBgMnAtsNDMEguDdyZ7MCLfVVi4k5kzZ9GvX7dUh1KjWCxGq5tv54jRv2P7iuXsWL2qZNnmubM5uPNJe62z9cMF1G59BGkH7XneZNf27awa/xCNf3jxXgm9JqvMmbOZPQ28AxxpZvlmdiVwj5l9aGYfAD2A6wHcfRHwLPETfa8A1wQz7CJgCPAqsBh4NhhbpvJqzlcBJ7r71uCM5SQzO9zdx1FG3T2xyP7nP99Bbu7F5cURKo0aNaCgYAM5OdkUFGwgO7tBqkOKnDffnEf79m343vcapjqUGimtTh3qtD2SrR8tJKtZc4q3bmXbyuU0z71mr7Gb582hXuc9zz95cRGr/vIw9bt05eBOJ1ZX2NWiMu9BcfdLSul+pIzxo4HRpfRPA6ZVZN/lzezT3H1rsPEVQHfgLDO7jzI+A3fPc/fO7t65piVmgJ49T+KFF2YA8MILM+jVq9wTq1JBU6e+yTnnnJHqMGqUoi1bKP7mGwB2FRby9ceLyWrcBIDN78+l7rEdiGVk7LFO8bZv+GbJxyUnCQHcnTVPPEZmk6Zk9+pTfQdQTWKWfAuz8mbOa82sk7vPBwhm0OcCE4Djqjy6ajBs2O+YPftDNm7czOmn/5Rrr/0Rubk/5LrrxjBp0nSaNj2EceOSuqFHkrRt23befns+d9yx9yxP9q1o8ybWPD4Bdu3C3al3QhfqHtcRgM3zZtOo99l7rbNl/vscdHR7YllZJX3bPl3K5tnvkNWsOcvv/A0Ah5w/kLrHdqieA6liYU+6yTL3fV/RYWYtgCJ3X1vKsm7u/p/yd/FJNJ58LQe0/q+tS3UIB4TJZ35/v1Prmm9eTDrnNK1zXmhTeZkzZ3fPL2NZEolZRKR66ZtQRERCKLRT4QpSchaRSInKI0OVnEUkUtJSHUAlUXIWkUjRzFlEJJSikZ2VnEUkUkzJWUQkfMyi8bBNJWcRiRjNnEVEQsci8ph6JWcRiRSVNUREQkllDRGR0NHVGiIiIaTkLCISQmbRuIFbyVlEIkYzZxGR0FFZQ0QklHQpnYhI6GjmLCISQhaRZ4YqOYtIpFhEHrev5CwiEaOZs4hI6KisISISSkrOIiKho0eGioiEkmbOIiKhE9PznEVEwkjJWUQkdHSHoIhIKCk5i4iEjq5zFhEJoajcvm3unuoYQsfMct09L9VxRJk+46qnz7hmi8ZpzcqXm+oADgD6jKuePuMaTMlZRCSElJxFREJIybl0qtNVPX3GVU+fcQ2mE4IiIiGkmbOISAgpOYuIhJCScwIz62dmH5vZUjMbkep4osjMJphZgZktTHUsUWVmLc3sdTNbbGaLzOyXqY5JKk4154CZpQGfAL2BfGAOcIm7f5TSwCLGzE4HtgKPu/uxqY4nisysKdDU3d8zs4OBecAA/VmuWTRz/tZJwFJ3X+buhcBEoH+KY4ocd38T2JDqOKLM3de4+3vB6y3AYqB5aqOSilJy/lZz4POE9/noD7TUcGZ2OHA8MCu1kUhFKTl/q7RHWanmIzWWmdUF/g5c5+6bUx2PVIyS87fygZYJ71sAq1MUi8h+MbMM4on5SXd/PtXxSMUpOX9rDtDWzFqZWSYwCJiS4phEKsziDzR+BFjs7velOh753yg5B9y9CBgCvEr8BMqz7r4otVFFj5k9DbwDHGlm+WZ2ZapjiqBuwE+AnmY2P2hnpzooqRhdSiciEkKaOYuIhJCSs4hICCk5i4iEkJKziEgIKTmLiISQkrOISAgpOYuIhND/Ax5/qcpfYgz1AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.heatmap(confusion_matrix(y_test, y_pred), cmap=\"YlGnBu\",annot=True, fmt=\"d\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       0.64      0.66      0.65      7731\n",
      "           2       0.68      0.65      0.66      8298\n",
      "           3       0.98      0.98      0.98      3782\n",
      "\n",
      "    accuracy                           0.72     19811\n",
      "   macro avg       0.77      0.77      0.77     19811\n",
      "weighted avg       0.72      0.72      0.72     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "adboost_with_pca = AdaBoostClassifier(n_estimators=400, random_state=93)\n",
    "adboost_with_pca.fit(pca.transform(X_train),y_train)\n",
    "y_pred = adboost_with_pca.predict(pca.transform(X_test))\n",
    "print(classification_report(y_pred,y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## neural network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_neural_network_model():\n",
    "    model = Sequential()\n",
    "\n",
    "    model.add(Dense(3, input_dim=len(X_train.columns), activation='relu'))\n",
    "    model.add(Dense(50, activation='relu'))\n",
    "    model.add(Dense(30, activation='relu'))\n",
    "    model.add(Dense(3, activation='softmax'))\n",
    "\n",
    "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "80.82%\n"
     ]
    }
   ],
   "source": [
    "np.random.seed(93)\n",
    "model = get_neural_network_model()\n",
    "model.fit(X_train, pd.get_dummies(y_train), epochs=10, validation_split=0.1,verbose=0)\n",
    "_, test_acc = model.evaluate(X_test, pd.get_dummies(y_test),verbose=0)\n",
    "print(\"%.2f%%\" % (test_acc*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       0.89      0.70      0.79     10114\n",
      "           2       0.63      0.85      0.72      5934\n",
      "           3       0.97      0.98      0.97      3763\n",
      "\n",
      "    accuracy                           0.80     19811\n",
      "   macro avg       0.83      0.84      0.83     19811\n",
      "weighted avg       0.83      0.80      0.80     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "y_pred = model.predict_classes(X_test)\n",
    "y_pred = pd.Series(y_pred)+1\n",
    "print(classification_report(y_pred,y_test))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bagging with Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "           1       0.97      0.98      0.98      8012\n",
      "           2       0.99      0.97      0.98      8017\n",
      "           3       0.99      1.00      0.99      3782\n",
      "\n",
      "    accuracy                           0.98     19811\n",
      "   macro avg       0.98      0.98      0.98     19811\n",
      "weighted avg       0.98      0.98      0.98     19811\n",
      "\n"
     ]
    }
   ],
   "source": [
    "rnd_clf = RandomForestClassifier(n_estimators=100,max_depth=100,n_jobs=-1)\n",
    "rnd_clf = rnd_clf.fit(X_train,y_train)\n",
    "y_pred = rnd_clf.predict(X_test)\n",
    "print(classification_report(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T13:17:48.394961Z",
     "start_time": "2020-07-31T13:17:13.755528Z"
    }
   },
   "outputs": [],
   "source": [
    "test_df = construct_datasets(\"Validation/*/*.csv\")\n",
    "test_df = test_df.dropna()\n",
    "y = test_df[\"state_type\"]\n",
    "X = test_df.drop([\"Frame ID\",\"state_type\",\"name\",\"Hand Type\",\"# hands\"], axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T13:17:57.993600Z",
     "start_time": "2020-07-31T13:17:57.988597Z"
    }
   },
   "outputs": [],
   "source": [
    "kfold = StratifiedKFold(n_splits=10, shuffle=True, random_state=93)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T13:19:34.857286Z",
     "start_time": "2020-07-31T13:19:34.852280Z"
    }
   },
   "outputs": [],
   "source": [
    "def print_score(cvscores):\n",
    "    print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.8816487848986263\n"
     ]
    }
   ],
   "source": [
    "cvscores = cross_val_score(knn, X, y, cv=kfold,n_jobs=-1)\n",
    "print_score(cvscores)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "54027/54027 [==============================] - 2s 39us/step - loss: 0.8985 - accuracy: 0.5901\n",
      "Epoch 2/10\n",
      "54027/54027 [==============================] - 3s 47us/step - loss: 0.5813 - accuracy: 0.7071\n",
      "Epoch 3/10\n",
      "54027/54027 [==============================] - 2s 43us/step - loss: 0.5051 - accuracy: 0.7734\n",
      "Epoch 4/10\n",
      "54027/54027 [==============================] - 3s 57us/step - loss: 0.4791 - accuracy: 0.7866\n",
      "Epoch 5/10\n",
      "54027/54027 [==============================] - 3s 55us/step - loss: 0.4639 - accuracy: 0.7944\n",
      "Epoch 6/10\n",
      "54027/54027 [==============================] - 2s 39us/step - loss: 0.4523 - accuracy: 0.8059\n",
      "Epoch 7/10\n",
      "54027/54027 [==============================] - 2s 46us/step - loss: 0.4467 - accuracy: 0.8057\n",
      "Epoch 8/10\n",
      " 2048/54027 [>.............................] - ETA: 1:03 - loss: 0.4357 - accuracy: 0.8062"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\magshimim\\Anaconda3\\lib\\site-packages\\keras\\callbacks\\callbacks.py:95: RuntimeWarning: Method (on_train_batch_end) is slow compared to the batch update (0.103001). Check your callbacks.\n",
      "  % (hook_name, delta_t_median), RuntimeWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "54027/54027 [==============================] - 8s 140us/step - loss: 0.4425 - accuracy: 0.8110\n",
      "Epoch 9/10\n",
      "54027/54027 [==============================] - 3s 54us/step - loss: 0.4414 - accuracy: 0.8099\n",
      "Epoch 10/10\n",
      "54027/54027 [==============================] - 2s 46us/step - loss: 0.4424 - accuracy: 0.8105\n",
      "accuracy: 80.40%\n",
      "Epoch 1/10\n",
      "54027/54027 [==============================] - 2s 46us/step - loss: 1.0359 - accuracy: 0.5384\n",
      "Epoch 2/10\n",
      "54027/54027 [==============================] - 3s 49us/step - loss: 0.6358 - accuracy: 0.6334\n",
      "Epoch 3/10\n",
      "54027/54027 [==============================] - 3s 47us/step - loss: 0.5887 - accuracy: 0.6921\n",
      "Epoch 4/10\n",
      "54027/54027 [==============================] - 3s 63us/step - loss: 0.5673 - accuracy: 0.7091\n",
      "Epoch 5/10\n",
      "54027/54027 [==============================] - 3s 60us/step - loss: 0.5470 - accuracy: 0.7267\n",
      "Epoch 6/10\n",
      "54027/54027 [==============================] - 3s 56us/step - loss: 0.5210 - accuracy: 0.7416\n",
      "Epoch 7/10\n",
      "54027/54027 [==============================] - 2s 39us/step - loss: 0.5137 - accuracy: 0.7488\n",
      "Epoch 8/10\n",
      "54027/54027 [==============================] - 2s 44us/step - loss: 0.5105 - accuracy: 0.7493\n",
      "Epoch 9/10\n",
      "54027/54027 [==============================] - 2s 38us/step - loss: 0.5043 - accuracy: 0.7533\n",
      "Epoch 10/10\n",
      "54027/54027 [==============================] - 2s 39us/step - loss: 0.4992 - accuracy: 0.7582\n",
      "accuracy: 75.49%\n",
      "Epoch 1/10\n",
      "54028/54028 [==============================] - 2s 42us/step - loss: 1.1197 - accuracy: 0.4411\n",
      "Epoch 2/10\n",
      "54028/54028 [==============================] - 2s 39us/step - loss: 0.8794 - accuracy: 0.5265\n",
      "Epoch 3/10\n",
      "54028/54028 [==============================] - 3s 47us/step - loss: 0.6206 - accuracy: 0.6915\n",
      "Epoch 4/10\n",
      "54028/54028 [==============================] - 2s 42us/step - loss: 0.5475 - accuracy: 0.7209\n",
      "Epoch 5/10\n",
      "54028/54028 [==============================] - 3s 50us/step - loss: 0.5220 - accuracy: 0.7362\n",
      "Epoch 6/10\n",
      "54028/54028 [==============================] - 3s 46us/step - loss: 0.5112 - accuracy: 0.7486\n",
      "Epoch 7/10\n",
      "54028/54028 [==============================] - 3s 51us/step - loss: 0.4967 - accuracy: 0.7561\n",
      "Epoch 8/10\n",
      "54028/54028 [==============================] - 3s 47us/step - loss: 0.4803 - accuracy: 0.7719\n",
      "Epoch 9/10\n",
      "54028/54028 [==============================] - 2s 42us/step - loss: 0.4609 - accuracy: 0.7846\n",
      "Epoch 10/10\n",
      "54028/54028 [==============================] - 3s 51us/step - loss: 0.4560 - accuracy: 0.7881\n",
      "accuracy: 78.96%\n",
      "Epoch 1/10\n",
      "54028/54028 [==============================] - 2s 42us/step - loss: 1.1087 - accuracy: 0.4465\n",
      "Epoch 2/10\n",
      "54028/54028 [==============================] - 2s 43us/step - loss: 0.7271 - accuracy: 0.6449\n",
      "Epoch 3/10\n",
      "54028/54028 [==============================] - 2s 37us/step - loss: 0.5447 - accuracy: 0.7478\n",
      "Epoch 4/10\n",
      "54028/54028 [==============================] - 2s 41us/step - loss: 0.4706 - accuracy: 0.7815\n",
      "Epoch 5/10\n",
      "54028/54028 [==============================] - 3s 52us/step - loss: 0.4549 - accuracy: 0.7900\n",
      "Epoch 6/10\n",
      "54028/54028 [==============================] - 3s 55us/step - loss: 0.4431 - accuracy: 0.7975\n",
      "Epoch 7/10\n",
      "54028/54028 [==============================] - 2s 44us/step - loss: 0.4363 - accuracy: 0.8000\n",
      "Epoch 8/10\n",
      "54028/54028 [==============================] - 2s 37us/step - loss: 0.4413 - accuracy: 0.7987\n",
      "Epoch 9/10\n",
      "54028/54028 [==============================] - 2s 46us/step - loss: 0.4393 - accuracy: 0.7996\n",
      "Epoch 10/10\n",
      "54028/54028 [==============================] - 2s 38us/step - loss: 0.4314 - accuracy: 0.8029\n",
      "accuracy: 81.45%\n",
      "Epoch 1/10\n",
      "54028/54028 [==============================] - 2s 41us/step - loss: 0.7791 - accuracy: 0.6147\n",
      "Epoch 2/10\n",
      "54028/54028 [==============================] - 3s 49us/step - loss: 0.4912 - accuracy: 0.7767\n",
      "Epoch 3/10\n",
      "54028/54028 [==============================] - 2s 44us/step - loss: 0.4714 - accuracy: 0.7894\n",
      "Epoch 4/10\n",
      "54028/54028 [==============================] - 3s 48us/step - loss: 0.4601 - accuracy: 0.7985\n",
      "Epoch 5/10\n",
      "54028/54028 [==============================] - 2s 40us/step - loss: 0.4459 - accuracy: 0.8081\n",
      "Epoch 6/10\n",
      "54028/54028 [==============================] - 3s 47us/step - loss: 0.4403 - accuracy: 0.8145\n",
      "Epoch 7/10\n",
      "54028/54028 [==============================] - 2s 45us/step - loss: 0.4341 - accuracy: 0.8172\n",
      "Epoch 8/10\n",
      "54028/54028 [==============================] - 2s 44us/step - loss: 0.4317 - accuracy: 0.8179\n",
      "Epoch 9/10\n",
      "54028/54028 [==============================] - 3s 52us/step - loss: 0.4290 - accuracy: 0.8172\n",
      "Epoch 10/10\n",
      "54028/54028 [==============================] - 2s 39us/step - loss: 0.4264 - accuracy: 0.8201\n",
      "accuracy: 80.56%\n",
      "Epoch 1/10\n",
      "54030/54030 [==============================] - 3s 46us/step - loss: 0.7113 - accuracy: 0.6090\n",
      "Epoch 2/10\n",
      "54030/54030 [==============================] - 2s 39us/step - loss: 0.5808 - accuracy: 0.6844\n",
      "Epoch 3/10\n",
      "54030/54030 [==============================] - 2s 46us/step - loss: 0.5298 - accuracy: 0.7371\n",
      "Epoch 4/10\n",
      "54030/54030 [==============================] - 2s 38us/step - loss: 0.5174 - accuracy: 0.7428\n",
      "Epoch 5/10\n",
      "54030/54030 [==============================] - ETA: 0s - loss: 0.5070 - accuracy: 0.74 - 2s 44us/step - loss: 0.5071 - accuracy: 0.7487\n",
      "Epoch 6/10\n",
      "54030/54030 [==============================] - 3s 47us/step - loss: 0.5035 - accuracy: 0.7474\n",
      "Epoch 7/10\n",
      "54030/54030 [==============================] - 3s 52us/step - loss: 0.5003 - accuracy: 0.7511\n",
      "Epoch 8/10\n",
      "54030/54030 [==============================] - 3s 53us/step - loss: 0.4962 - accuracy: 0.7525\n",
      "Epoch 9/10\n",
      "54030/54030 [==============================] - 2s 43us/step - loss: 0.4936 - accuracy: 0.7539\n",
      "Epoch 10/10\n",
      "54030/54030 [==============================] - 3s 50us/step - loss: 0.4942 - accuracy: 0.7571\n",
      "accuracy: 75.12%\n",
      "Epoch 1/10\n",
      "54030/54030 [==============================] - 2s 45us/step - loss: 0.7603 - accuracy: 0.6319\n",
      "Epoch 2/10\n",
      "54030/54030 [==============================] - 3s 50us/step - loss: 0.5706 - accuracy: 0.7140\n",
      "Epoch 3/10\n",
      "54030/54030 [==============================] - 2s 45us/step - loss: 0.5368 - accuracy: 0.7314\n",
      "Epoch 4/10\n",
      "54030/54030 [==============================] - 3s 50us/step - loss: 0.5148 - accuracy: 0.7388\n",
      "Epoch 5/10\n",
      "54030/54030 [==============================] - 2s 43us/step - loss: 0.5073 - accuracy: 0.7460\n",
      "Epoch 6/10\n",
      "54030/54030 [==============================] - 3s 54us/step - loss: 0.5002 - accuracy: 0.7486\n",
      "Epoch 7/10\n",
      "54030/54030 [==============================] - 2s 44us/step - loss: 0.5002 - accuracy: 0.7465\n",
      "Epoch 8/10\n",
      "54030/54030 [==============================] - 2s 43us/step - loss: 0.4927 - accuracy: 0.7512\n",
      "Epoch 9/10\n",
      "54030/54030 [==============================] - 3s 48us/step - loss: 0.4933 - accuracy: 0.7508\n",
      "Epoch 10/10\n",
      "54030/54030 [==============================] - 2s 42us/step - loss: 0.4910 - accuracy: 0.7530\n",
      "accuracy: 76.07%\n",
      "Epoch 1/10\n",
      "54030/54030 [==============================] - 3s 51us/step - loss: 0.9207 - accuracy: 0.6033\n",
      "Epoch 2/10\n",
      "54030/54030 [==============================] - 2s 43us/step - loss: 0.5817 - accuracy: 0.6712\n",
      "Epoch 3/10\n",
      "54030/54030 [==============================] - 3s 49us/step - loss: 0.5457 - accuracy: 0.6874\n",
      "Epoch 4/10\n",
      "54030/54030 [==============================] - 2s 41us/step - loss: 0.5305 - accuracy: 0.7235\n",
      "Epoch 5/10\n",
      "54030/54030 [==============================] - 3s 49us/step - loss: 0.5146 - accuracy: 0.7505\n",
      "Epoch 6/10\n",
      "54030/54030 [==============================] - 2s 42us/step - loss: 0.4931 - accuracy: 0.7731\n",
      "Epoch 7/10\n",
      "54030/54030 [==============================] - 2s 41us/step - loss: 0.4852 - accuracy: 0.7809\n",
      "Epoch 8/10\n",
      "54030/54030 [==============================] - 3s 48us/step - loss: 0.4744 - accuracy: 0.7894\n",
      "Epoch 9/10\n",
      "54030/54030 [==============================] - 2s 43us/step - loss: 0.4745 - accuracy: 0.7908\n",
      "Epoch 10/10\n",
      "54030/54030 [==============================] - 3s 56us/step - loss: 0.4655 - accuracy: 0.7976\n",
      "accuracy: 78.54%\n",
      "Epoch 1/10\n",
      "54030/54030 [==============================] - 3s 47us/step - loss: 0.9724 - accuracy: 0.5732\n",
      "Epoch 2/10\n",
      "54030/54030 [==============================] - 3s 54us/step - loss: 0.5679 - accuracy: 0.6933\n",
      "Epoch 3/10\n",
      "54030/54030 [==============================] - 3s 47us/step - loss: 0.5142 - accuracy: 0.7388\n",
      "Epoch 4/10\n",
      "54030/54030 [==============================] - 3s 62us/step - loss: 0.4913 - accuracy: 0.7531\n",
      "Epoch 5/10\n",
      "54030/54030 [==============================] - 3s 55us/step - loss: 0.4664 - accuracy: 0.7671\n",
      "Epoch 6/10\n",
      "54030/54030 [==============================] - 4s 65us/step - loss: 0.4534 - accuracy: 0.7761\n",
      "Epoch 7/10\n",
      "54030/54030 [==============================] - 4s 68us/step - loss: 0.4389 - accuracy: 0.7854\n",
      "Epoch 8/10\n",
      "54030/54030 [==============================] - 3s 55us/step - loss: 0.4328 - accuracy: 0.7896\n",
      "Epoch 9/10\n",
      "54030/54030 [==============================] - 3s 51us/step - loss: 0.4186 - accuracy: 0.7979\n",
      "Epoch 10/10\n",
      "54030/54030 [==============================] - 3s 49us/step - loss: 0.4138 - accuracy: 0.8016\n",
      "accuracy: 80.07%\n",
      "Epoch 1/10\n",
      "54030/54030 [==============================] - 3s 55us/step - loss: 0.9424 - accuracy: 0.5808\n",
      "Epoch 2/10\n",
      "54030/54030 [==============================] - 3s 48us/step - loss: 0.6138 - accuracy: 0.6464\n",
      "Epoch 3/10\n",
      "54030/54030 [==============================] - 3s 55us/step - loss: 0.5734 - accuracy: 0.7016\n",
      "Epoch 4/10\n",
      "54030/54030 [==============================] - 2s 41us/step - loss: 0.5416 - accuracy: 0.7341\n",
      "Epoch 5/10\n",
      "54030/54030 [==============================] - 3s 49us/step - loss: 0.5124 - accuracy: 0.7623\n",
      "Epoch 6/10\n",
      "54030/54030 [==============================] - 2s 39us/step - loss: 0.4648 - accuracy: 0.7913\n",
      "Epoch 7/10\n",
      "54030/54030 [==============================] - 2s 41us/step - loss: 0.4316 - accuracy: 0.8095\n",
      "Epoch 8/10\n",
      "54030/54030 [==============================] - 3s 51us/step - loss: 0.4214 - accuracy: 0.8143\n",
      "Epoch 9/10\n",
      "54030/54030 [==============================] - 2s 44us/step - loss: 0.4132 - accuracy: 0.8217\n",
      "Epoch 10/10\n",
      "54030/54030 [==============================] - 2s 46us/step - loss: 0.4092 - accuracy: 0.8221\n",
      "accuracy: 83.37%\n",
      "79.00% (+/- 2.58%)\n"
     ]
    }
   ],
   "source": [
    "cvscores = []\n",
    "for train, test in kfold.split(X, y):\n",
    "    model = get_neural_network_model()\n",
    "    model.fit(X.iloc[train], pd.get_dummies(y[train]), epochs=10,verbose=0)\n",
    "    scores = model.evaluate(X.iloc[test], pd.get_dummies(y[test]), verbose=0)\n",
    "    cvscores.append(scores[1] * 100)\n",
    "print(cvscores)\n",
    "print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0]\ttrain-merror:0.12371\tvalidation-merror:0.12377\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00076\tvalidation-merror:0.00453\n",
      "[0]\ttrain-merror:0.12618\tvalidation-merror:0.12725\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00076\tvalidation-merror:0.00506\n",
      "[0]\ttrain-merror:0.12792\tvalidation-merror:0.12778\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00063\tvalidation-merror:0.00346\n",
      "[0]\ttrain-merror:0.10663\tvalidation-merror:0.11847\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00095\tvalidation-merror:0.00533\n",
      "[0]\ttrain-merror:0.12144\tvalidation-merror:0.12435\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00093\tvalidation-merror:0.00507\n",
      "[0]\ttrain-merror:0.11942\tvalidation-merror:0.12582\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00070\tvalidation-merror:0.00520\n",
      "[0]\ttrain-merror:0.12142\tvalidation-merror:0.12968\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00088\tvalidation-merror:0.00453\n",
      "[0]\ttrain-merror:0.12467\tvalidation-merror:0.13941\n",
      "Multiple eval metrics have been passed: 'validation-merror' will be used for early stopping.\n",
      "\n",
      "Will train until validation-merror hasn't improved in 50 rounds.\n",
      "[299]\ttrain-merror:0.00072\tvalidation-merror:0.00520\n",
      "1.00% (+/- 0.00%)\n"
     ]
    }
   ],
   "source": [
    "kfold = StratifiedKFold(n_splits=8, shuffle=True, random_state=93)\n",
    "cvscores = []\n",
    "for train, test in kfold.split(X, y):\n",
    "    y_pred = xgboost_model(X_train=X.iloc[train],X_test=X.iloc[test],y_train=y[train],y_test=y[test],verbose_eval=600)\n",
    "    cvscores.append(accuracy_score(y[test],y_pred)*100)\n",
    "print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.985724153548933\n"
     ]
    }
   ],
   "source": [
    "cvscores = cross_val_score(rnd_clf, X, y, cv=kfold,n_jobs=-1)*100\n",
    "print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2020-07-31T14:58:00.625643Z",
     "start_time": "2020-07-31T13:20:31.580231Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "99.43% (+/- 0.35%)\n"
     ]
    }
   ],
   "source": [
    "cvscores = cross_val_score(adboost, X, y, cv=kfold,n_jobs=-1)*100\n",
    "print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "94.28% (+/- 0.32%)\n"
     ]
    }
   ],
   "source": [
    "cvscores = cross_val_score(dst, X, y, cv=kfold,n_jobs=-1)*100\n",
    "print(\"%.2f%% (+/- %.2f%%)\" % (np.mean(cvscores), np.std(cvscores)))"
   ]
  },
  {
   "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": "3.7.4"
  },
  "notify_time": "5",
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}