Model - Copy.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8f37bb9d",
   "metadata": {},
   "source": [
    "# Importing the Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "df344e92",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "7af38d43",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "from sklearn.model_selection import train_test_split\n",
    "from pandas import DataFrame\n",
    "from pandas import concat"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "3e8b0ba5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import functions as f"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bdab091c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "5c6df7a7",
   "metadata": {},
   "source": [
    "### Functions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c426cd60",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "0289e3e0",
   "metadata": {},
   "source": [
    "## Preparing the Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "f63452eb",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv(\"datasets/csvs-old/Seq_ETH.csv\")\n",
    "pos_df = df.drop(['v_x', 'v_y'], axis = 1)\n",
    "pos_grp = pos_df.groupby('pedestrian_ID')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "e7834653",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>frame_number</th>\n",
       "      <th>pedestrian_ID</th>\n",
       "      <th>pos_x</th>\n",
       "      <th>pos_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>77</td>\n",
       "      <td>1</td>\n",
       "      <td>8.456844</td>\n",
       "      <td>3.588066</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>78</td>\n",
       "      <td>1</td>\n",
       "      <td>9.125530</td>\n",
       "      <td>3.658583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>79</td>\n",
       "      <td>1</td>\n",
       "      <td>9.787146</td>\n",
       "      <td>3.849445</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>79</td>\n",
       "      <td>1</td>\n",
       "      <td>10.472197</td>\n",
       "      <td>3.955450</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>80</td>\n",
       "      <td>1</td>\n",
       "      <td>11.066000</td>\n",
       "      <td>4.061280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8903</th>\n",
       "      <td>1238</td>\n",
       "      <td>357</td>\n",
       "      <td>10.449366</td>\n",
       "      <td>6.229233</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8904</th>\n",
       "      <td>1238</td>\n",
       "      <td>367</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8905</th>\n",
       "      <td>1238</td>\n",
       "      <td>366</td>\n",
       "      <td>12.132311</td>\n",
       "      <td>6.984982</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8906</th>\n",
       "      <td>1238</td>\n",
       "      <td>364</td>\n",
       "      <td>12.369087</td>\n",
       "      <td>4.146375</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8907</th>\n",
       "      <td>1238</td>\n",
       "      <td>365</td>\n",
       "      <td>12.708071</td>\n",
       "      <td>5.336541</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>8908 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      frame_number  pedestrian_ID      pos_x     pos_y\n",
       "0               77              1   8.456844  3.588066\n",
       "1               78              1   9.125530  3.658583\n",
       "2               79              1   9.787146  3.849445\n",
       "3               79              1  10.472197  3.955450\n",
       "4               80              1  11.066000  4.061280\n",
       "...            ...            ...        ...       ...\n",
       "8903          1238            357  10.449366  6.229233\n",
       "8904          1238            367  11.201661  8.443910\n",
       "8905          1238            366  12.132311  6.984982\n",
       "8906          1238            364  12.369087  4.146375\n",
       "8907          1238            365  12.708071  5.336541\n",
       "\n",
       "[8908 rows x 4 columns]"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "c7617fd9",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "17\n",
      "360\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "length = []\n",
    "for name, group in pos_grp:\n",
    "    length.append(len(group))\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.hist(length, bins = 100)\n",
    "\n",
    "newL = [x for x in length if x < 9]\n",
    "print(len(newL))\n",
    "print(len(length))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "66817c19",
   "metadata": {},
   "outputs": [],
   "source": [
    "df1 = df.copy()\n",
    "for name, group in pos_grp:\n",
    "    if len(group) < 9:\n",
    "        df1 = df1.drop(pos_grp.get_group(name).index)\n",
    "\n",
    "pos_df1 = df1.drop(['v_x', 'v_y'], axis = 1)\n",
    "pos_grp1 = pos_df1.groupby('pedestrian_ID')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "1e4b513c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0\n",
      "343\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "length = []\n",
    "for name, group in pos_grp1:\n",
    "    length.append(len(group))\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.hist(length, bins = 100)\n",
    "\n",
    "# newL = [x for x in length if x < 7]\n",
    "# print(len(newL))\n",
    "# print(len(length))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "f6664d35",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_lstm = pd.DataFrame()\n",
    "for name, group in pos_grp1:\n",
    "    first_cols = group[['frame_number', 'pedestrian_ID']]\n",
    "    second_cols = group[['pos_x', 'pos_y']]\n",
    "    second_cols_converted = f.series_to_supervised(second_cols, 4, 4)\n",
    "    \n",
    "    # df_temp = pd.concat([first_cols, second_cols_converted], axis=1)\n",
    "    df_lstm = pd.concat([df_lstm, second_cols_converted])\n",
    "    \n",
    "df_lstm = df_lstm.dropna()\n",
    "# df_lstm = df_lstm.drop(['frame_number', 'pedestrian_ID'], axis = 1)\n",
    "# df_lstm.to_csv(f\"datasets/csvs/lstms/seq_eth.csv\", index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0f1206bf",
   "metadata": {},
   "source": [
    "## Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "241548c4",
   "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>var1(t-4)</th>\n",
       "      <th>var2(t-4)</th>\n",
       "      <th>var1(t-3)</th>\n",
       "      <th>var2(t-3)</th>\n",
       "      <th>var1(t-2)</th>\n",
       "      <th>var2(t-2)</th>\n",
       "      <th>var1(t-1)</th>\n",
       "      <th>var2(t-1)</th>\n",
       "      <th>var1(t)</th>\n",
       "      <th>var2(t)</th>\n",
       "      <th>var1(t+1)</th>\n",
       "      <th>var2(t+1)</th>\n",
       "      <th>var1(t+2)</th>\n",
       "      <th>var2(t+2)</th>\n",
       "      <th>var1(t+3)</th>\n",
       "      <th>var2(t+3)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>13.017548</td>\n",
       "      <td>5.782591</td>\n",
       "      <td>12.087770</td>\n",
       "      <td>5.751949</td>\n",
       "      <td>11.746306</td>\n",
       "      <td>5.729815</td>\n",
       "      <td>11.175158</td>\n",
       "      <td>5.836247</td>\n",
       "      <td>10.434290</td>\n",
       "      <td>5.873203</td>\n",
       "      <td>10.050254</td>\n",
       "      <td>6.171487</td>\n",
       "      <td>9.571296</td>\n",
       "      <td>6.237355</td>\n",
       "      <td>9.084074</td>\n",
       "      <td>6.263836</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>12.087770</td>\n",
       "      <td>5.751949</td>\n",
       "      <td>11.746306</td>\n",
       "      <td>5.729815</td>\n",
       "      <td>11.175158</td>\n",
       "      <td>5.836247</td>\n",
       "      <td>10.434290</td>\n",
       "      <td>5.873203</td>\n",
       "      <td>10.050254</td>\n",
       "      <td>6.171487</td>\n",
       "      <td>9.571296</td>\n",
       "      <td>6.237355</td>\n",
       "      <td>9.084074</td>\n",
       "      <td>6.263836</td>\n",
       "      <td>8.552751</td>\n",
       "      <td>6.374027</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>11.746306</td>\n",
       "      <td>5.729815</td>\n",
       "      <td>11.175158</td>\n",
       "      <td>5.836247</td>\n",
       "      <td>10.434290</td>\n",
       "      <td>5.873203</td>\n",
       "      <td>10.050254</td>\n",
       "      <td>6.171487</td>\n",
       "      <td>9.571296</td>\n",
       "      <td>6.237355</td>\n",
       "      <td>9.084074</td>\n",
       "      <td>6.263836</td>\n",
       "      <td>8.552751</td>\n",
       "      <td>6.374027</td>\n",
       "      <td>8.097887</td>\n",
       "      <td>6.481037</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>11.175158</td>\n",
       "      <td>5.836247</td>\n",
       "      <td>10.434290</td>\n",
       "      <td>5.873203</td>\n",
       "      <td>10.050254</td>\n",
       "      <td>6.171487</td>\n",
       "      <td>9.571296</td>\n",
       "      <td>6.237355</td>\n",
       "      <td>9.084074</td>\n",
       "      <td>6.263836</td>\n",
       "      <td>8.552751</td>\n",
       "      <td>6.374027</td>\n",
       "      <td>8.097887</td>\n",
       "      <td>6.481037</td>\n",
       "      <td>7.635367</td>\n",
       "      <td>6.548341</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>10.434290</td>\n",
       "      <td>5.873203</td>\n",
       "      <td>10.050254</td>\n",
       "      <td>6.171487</td>\n",
       "      <td>9.571296</td>\n",
       "      <td>6.237355</td>\n",
       "      <td>9.084074</td>\n",
       "      <td>6.263836</td>\n",
       "      <td>8.552751</td>\n",
       "      <td>6.374027</td>\n",
       "      <td>8.097887</td>\n",
       "      <td>6.481037</td>\n",
       "      <td>7.635367</td>\n",
       "      <td>6.548341</td>\n",
       "      <td>7.167091</td>\n",
       "      <td>6.616482</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8856</th>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8864</th>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8872</th>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8879</th>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8886</th>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "      <td>11.201661</td>\n",
       "      <td>8.443910</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>6431 rows × 16 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      var1(t-4)  var2(t-4)  var1(t-3)  var2(t-3)  var1(t-2)  var2(t-2)  \\\n",
       "11    13.017548   5.782591  12.087770   5.751949  11.746306   5.729815   \n",
       "13    12.087770   5.751949  11.746306   5.729815  11.175158   5.836247   \n",
       "15    11.746306   5.729815  11.175158   5.836247  10.434290   5.873203   \n",
       "19    11.175158   5.836247  10.434290   5.873203  10.050254   6.171487   \n",
       "24    10.434290   5.873203  10.050254   6.171487   9.571296   6.237355   \n",
       "...         ...        ...        ...        ...        ...        ...   \n",
       "8856  11.201661   8.443910  11.201661   8.443910  11.201661   8.443910   \n",
       "8864  11.201661   8.443910  11.201661   8.443910  11.201661   8.443910   \n",
       "8872  11.201661   8.443910  11.201661   8.443910  11.201661   8.443910   \n",
       "8879  11.201661   8.443910  11.201661   8.443910  11.201661   8.443910   \n",
       "8886  11.201661   8.443910  11.201661   8.443910  11.201661   8.443910   \n",
       "\n",
       "      var1(t-1)  var2(t-1)    var1(t)   var2(t)  var1(t+1)  var2(t+1)  \\\n",
       "11    11.175158   5.836247  10.434290  5.873203  10.050254   6.171487   \n",
       "13    10.434290   5.873203  10.050254  6.171487   9.571296   6.237355   \n",
       "15    10.050254   6.171487   9.571296  6.237355   9.084074   6.263836   \n",
       "19     9.571296   6.237355   9.084074  6.263836   8.552751   6.374027   \n",
       "24     9.084074   6.263836   8.552751  6.374027   8.097887   6.481037   \n",
       "...         ...        ...        ...       ...        ...        ...   \n",
       "8856  11.201661   8.443910  11.201661  8.443910  11.201661   8.443910   \n",
       "8864  11.201661   8.443910  11.201661  8.443910  11.201661   8.443910   \n",
       "8872  11.201661   8.443910  11.201661  8.443910  11.201661   8.443910   \n",
       "8879  11.201661   8.443910  11.201661  8.443910  11.201661   8.443910   \n",
       "8886  11.201661   8.443910  11.201661  8.443910  11.201661   8.443910   \n",
       "\n",
       "      var1(t+2)  var2(t+2)  var1(t+3)  var2(t+3)  \n",
       "11     9.571296   6.237355   9.084074   6.263836  \n",
       "13     9.084074   6.263836   8.552751   6.374027  \n",
       "15     8.552751   6.374027   8.097887   6.481037  \n",
       "19     8.097887   6.481037   7.635367   6.548341  \n",
       "24     7.635367   6.548341   7.167091   6.616482  \n",
       "...         ...        ...        ...        ...  \n",
       "8856  11.201661   8.443910  11.201661   8.443910  \n",
       "8864  11.201661   8.443910  11.201661   8.443910  \n",
       "8872  11.201661   8.443910  11.201661   8.443910  \n",
       "8879  11.201661   8.443910  11.201661   8.443910  \n",
       "8886  11.201661   8.443910  11.201661   8.443910  \n",
       "\n",
       "[6431 rows x 16 columns]"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_lstm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "2fa416c7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(4501, 1, 8) (4501, 8) (1930, 1, 8) (1930, 8)\n"
     ]
    }
   ],
   "source": [
    "X = df_lstm.iloc[:,:8].values\n",
    "y = df_lstm.iloc[:,8:].values\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=4)\n",
    "\n",
    "# reshape input to be 3D [samples, timesteps, features]\n",
    "X_train = X_train.reshape((X_train.shape[0], 1, X_train.shape[1]))\n",
    "X_test = X_test.reshape((X_test.shape[0], 1, X_test.shape[1]))\n",
    "print(X_train.shape, y_train.shape, X_test.shape, y_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "191dfccb",
   "metadata": {},
   "outputs": [],
   "source": [
    "n_timesteps, n_features, n_outputs = X_train.shape[1], X_train.shape[2], y_train.shape[1]\n",
    "verbose, epochs, batch_size = 2, 50, 16"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "0abcc34d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_10\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm_20 (LSTM)               (None, 1, 32)             5248      \n",
      "_________________________________________________________________\n",
      "lstm_21 (LSTM)               (None, 32)                8320      \n",
      "_________________________________________________________________\n",
      "dense_10 (Dense)             (None, 8)                 264       \n",
      "=================================================================\n",
      "Total params: 13,832\n",
      "Trainable params: 13,832\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model1 = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.LSTM(32, return_sequences=True, input_shape=(X_train.shape[1], X_train.shape[2])),\n",
    "    tf.keras.layers.LSTM(32),\n",
    "\n",
    "    tf.keras.layers.Dense(8)\n",
    "])\n",
    "model1.summary()\n",
    "\n",
    "# model1.compile(loss='mae', optimizer='adam', metrics='acc')\n",
    "\n",
    "# history1 = model1.fit(X_train, y_train, epochs=5, batch_size=batch_size, validation_data=(X_test, y_test), verbose=verbose, shuffle=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "ebe2f8b5",
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "282/282 - 22s - loss: 2.8199 - acc: 0.2862 - val_loss: 1.3367 - val_acc: 0.4933\n",
      "Epoch 2/5\n",
      "282/282 - 6s - loss: 0.9603 - acc: 0.5803 - val_loss: 0.6811 - val_acc: 0.6212\n",
      "Epoch 3/5\n",
      "282/282 - 7s - loss: 0.5270 - acc: 0.6519 - val_loss: 0.3874 - val_acc: 0.6554\n",
      "Epoch 4/5\n",
      "282/282 - 7s - loss: 0.3484 - acc: 0.6645 - val_loss: 0.3143 - val_acc: 0.6731\n",
      "Epoch 5/5\n",
      "282/282 - 6s - loss: 0.2759 - acc: 0.6807 - val_loss: 0.2521 - val_acc: 0.6927\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.LSTM(32, return_sequences=True),\n",
    "    tf.keras.layers.LSTM(32),\n",
    "\n",
    "    tf.keras.layers.Dense(8)\n",
    "])\n",
    "\n",
    "model.compile(loss='mae', optimizer='adam', metrics='acc')\n",
    "\n",
    "history = model.fit(X_train, y_train, epochs=5, batch_size=batch_size, validation_data=(X_test, y_test), verbose=verbose, shuffle=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "18147876",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\Users\\lauki\\anaconda3\\envs\\tfgpu\\lib\\site-packages\\seaborn\\_decorators.py:36: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n",
      "c:\\Users\\lauki\\anaconda3\\envs\\tfgpu\\lib\\site-packages\\seaborn\\_decorators.py:36: FutureWarning: Pass the following variable as a keyword arg: x. From version 0.12, the only valid positional argument will be `data`, and passing other arguments without an explicit keyword will result in an error or misinterpretation.\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.lineplot(history.history['loss'], label='train')\n",
    "sns.lineplot(history.history['val_loss'], label='test')\n",
    "# plt.legend()\n",
    "sns.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "65d6b2c3",
   "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>loss</th>\n",
       "      <th>val_loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2.819928</td>\n",
       "      <td>1.336726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.960289</td>\n",
       "      <td>0.681055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.527029</td>\n",
       "      <td>0.387413</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.348376</td>\n",
       "      <td>0.314335</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.275950</td>\n",
       "      <td>0.252135</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       loss  val_loss\n",
       "0  2.819928  1.336726\n",
       "1  0.960289  0.681055\n",
       "2  0.527029  0.387413\n",
       "3  0.348376  0.314335\n",
       "4  0.275950  0.252135"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(history.history)[['loss', 'val_loss']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "01701de2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.lineplot(data=pd.DataFrame(history.history)[['loss', 'val_loss']])\n",
    "plt.show()\n",
    "sns.lineplot(data=pd.DataFrame(history.history)[['acc', 'val_acc']])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "298f87de",
   "metadata": {},
   "source": [
    "### Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "1dff7961",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.03322618 1.26998856 1.47109964 ... 2.38018962 2.44699327 1.41535632]\n",
      " [1.28255814 0.11970422 0.4587997  ... 3.66221888 3.72902253 2.69738558]\n",
      " [1.4629079  0.56398943 0.01451448 ... 3.84256865 3.9093723  2.87773535]\n",
      " ...\n",
      " [2.39573989 3.66625733 3.86736841 ... 0.03502897 0.08987312 2.17478083]\n",
      " [2.44784621 3.71836365 3.91947473 ... 0.06818547 0.03894446 2.22570949]\n",
      " [1.38084517 2.6513626  2.85247369 ... 2.20235184 2.25719599 0.03398228]]\n"
     ]
    }
   ],
   "source": [
    "y_pred = []\n",
    "for val in X_test:\n",
    "    pred = model.predict(val[np.newaxis])\n",
    "    y_pred.append(pred)\n",
    "\n",
    "print(tf.keras.metrics.mean_absolute_error(y_pred, y_test).numpy())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "281821f6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "-0.055778347"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred[0][0][0]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5305bd1f",
   "metadata": {},
   "source": [
    "#### Error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "id": "60e67c47",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.0323954"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "error = tf.keras.metrics.MeanAbsoluteError()\n",
    "error.update_state(y_pred, y_test)\n",
    "error.result().numpy()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "68f7134d",
   "metadata": {},
   "source": [
    "## Forcasting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "id": "4ace9f1e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_forecast(model, df, size_of_prediction=2, window_size=3):\n",
    "  \"\"\"\n",
    "  Makes the forecast on the data\n",
    "\n",
    "  Args:\n",
    "    model (tensorflow sequential) -- the trained model with 2 output tensors\n",
    "    df (pandas dataframe) -- the part of the dataframe on which the predictions are to be made\n",
    "    size_of_prediction (int) -- dimension of the prediction\n",
    "    window_size (int) -- size of the window\n",
    "  \"\"\"\n",
    "\n",
    "  series = []\n",
    "  forecasts = []\n",
    "  len_of_df = len(df)\n",
    "\n",
    "  try:\n",
    "    for i in range(window_size):\n",
    "      for j in range(size_of_prediction):\n",
    "        series.append(df.iloc[i][j])\n",
    "  except IndexError:\n",
    "    print(f\"Length of the dataframe={len(df)} is smaller than the window_size={window_size}. Add more data or reduce the window_size\")\n",
    "\n",
    "  for i in range(len_of_df-window_size):\n",
    "    predict = np.array(series[-window_size*size_of_prediction:])[np.newaxis]\n",
    "    forecast = model.predict(predict[-window_size*size_of_prediction:][np.newaxis])\n",
    "    forecasts.append(forecast[0])\n",
    "    for j in range(size_of_prediction):\n",
    "      series.append(forecast[0][j])\n",
    "\n",
    "  return np.array(forecasts)\n",
    "      "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "id": "a1b808a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_predictions(df, forecasts):\n",
    "    \"\"\"\n",
    "    Draws the predictions of the data\n",
    "\n",
    "    Args:\n",
    "    df (pandas dataframe) -- the part of the dataframe on which the predictions are to be made\n",
    "    forecasts (array of int) -- all the predictions from make_forecast()\n",
    "    \"\"\"\n",
    "    x_val = [val[0] for val in forecasts]\n",
    "    y_val = [val[1] for val in forecasts]\n",
    "\n",
    "    b = plt.scatter(np.array(df.iloc[:3,0]), np.array(df.iloc[:3,1]), c='b')\n",
    "    c = plt.scatter(np.array(df.iloc[3:,0]), np.array(df.iloc[3:,1]), c='c')\n",
    "    o = plt.scatter(x_val, y_val, c='orange')\n",
    "    plt.legend((b, c, o), ('before', 'after', 'prediction'))\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3226dba8",
   "metadata": {},
   "outputs": [],
   "source": [
    "i = 15\n",
    "for name, group in pos_grp1:\n",
    "    group = group.iloc[:,-2:]\n",
    "    forecasts = make_forecast(model, group)\n",
    "    draw_predictions(group, forecasts)\n",
    "    print(group)\n",
    "    i -= 1\n",
    "    if not i:\n",
    "        break"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3ecae1ed",
   "metadata": {},
   "source": [
    "# TEST"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "id": "129a2b97",
   "metadata": {},
   "outputs": [],
   "source": [
    "def future_forecast(model, df, size_of_prediction=2, window_size=3, no_of_forecasts=3):\n",
    "  \"\"\"\n",
    "  Makes the forecast on the data\n",
    "\n",
    "  Args:\n",
    "    model (tensorflow sequential) -- the trained model with 2 output tensors\n",
    "    df (pandas dataframe) -- the part of the dataframe on which the predictions are to be made\n",
    "    dim (int) -- dimension of the prediction\n",
    "    window_size (int) -- size of the window\n",
    "    no_of_forecasts (int) -- number of forecasts to be made, the forecasts made would also be used to make the later forecasts\n",
    "  \"\"\"\n",
    "\n",
    "  series = []\n",
    "  forecasts = []\n",
    "  len_of_df = len(df)\n",
    "\n",
    "  try:\n",
    "    for i in range(window_size):\n",
    "      for j in range(size_of_prediction):\n",
    "        series.append(df.iloc[i][j])\n",
    "  except IndexError:\n",
    "    print(f\"Length of the dataframe={len(df)} is smaller than the window_size={window_size}. Add more data or reduce the window_size\")\n",
    "\n",
    "  for i in range(no_of_forecasts):\n",
    "    predict = np.array(series[-window_size*size_of_prediction:])[np.newaxis]\n",
    "    forecast = model.predict(predict[-window_size*size_of_prediction:][np.newaxis])\n",
    "    forecasts.append(forecast[0])\n",
    "    for j in range(size_of_prediction):\n",
    "      series.append(forecast[0][j])\n",
    "\n",
    "  return np.array(forecasts)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "id": "259ae587",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.6184893, 2.9211946],\n",
       "       [0.6327673, 3.2772484],\n",
       "       [0.6006485, 3.5886047]], dtype=float32)"
      ]
     },
     "execution_count": 192,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "future_forecast(model, group)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "id": "46af7a72",
   "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>pos_x</th>\n",
       "      <th>pos_y</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>223</th>\n",
       "      <td>0.364292</td>\n",
       "      <td>1.678611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>226</th>\n",
       "      <td>0.445974</td>\n",
       "      <td>2.107313</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>229</th>\n",
       "      <td>0.585037</td>\n",
       "      <td>2.545277</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>232</th>\n",
       "      <td>0.611068</td>\n",
       "      <td>3.001580</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        pos_x     pos_y\n",
       "223  0.364292  1.678611\n",
       "226  0.445974  2.107313\n",
       "229  0.585037  2.545277\n",
       "232  0.611068  3.001580"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "group"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "38a034f0438b23d2a1aa6bdbbcadf64caf15722bd082d1ef34a2dbf1d449835e"
  },
  "kernelspec": {
   "display_name": "Python 3.9.12 ('tfgpu')",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}