simple_baseline_with_transform.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.linear_model import LinearRegression, RidgeCV, Ridge\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from coosys import cartesian_to_kepler as ctk, cartesian_to_quaternion as ctq\n",
    "from coosys import kepler_to_cartesian as ktc, quaternion_to_cartesian as qtc\n",
    "\n",
    "from sklearn.preprocessing import PolynomialFeatures as PF\n",
    "from tqdm.auto import tqdm\n",
    "tqdm.pandas()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Train simple baseline (Linear Regression models)\n",
    "\n",
    "metric: [MAPE](https://en.wikipedia.org/wiki/Mean_absolute_percentage_error)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(649912, 15)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load data\n",
    "train = pd.read_csv('train.csv')\n",
    "test = pd.read_csv('Track 1/test.csv')\n",
    "train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>id</th>\n",
       "      <th>sat_id</th>\n",
       "      <th>epoch</th>\n",
       "      <th>x_sim</th>\n",
       "      <th>y_sim</th>\n",
       "      <th>z_sim</th>\n",
       "      <th>Vx_sim</th>\n",
       "      <th>Vy_sim</th>\n",
       "      <th>Vz_sim</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3927</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:01:45.162</td>\n",
       "      <td>-13366.891347</td>\n",
       "      <td>-14236.753503</td>\n",
       "      <td>6386.774555</td>\n",
       "      <td>4.333815</td>\n",
       "      <td>-0.692764</td>\n",
       "      <td>0.810774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3928</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:22:57.007</td>\n",
       "      <td>-7370.434039</td>\n",
       "      <td>-14498.771520</td>\n",
       "      <td>7130.411325</td>\n",
       "      <td>5.077413</td>\n",
       "      <td>0.360609</td>\n",
       "      <td>0.313402</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3929</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:44:08.852</td>\n",
       "      <td>-572.068654</td>\n",
       "      <td>-13065.289498</td>\n",
       "      <td>7033.794876</td>\n",
       "      <td>5.519106</td>\n",
       "      <td>2.012830</td>\n",
       "      <td>-0.539412</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3930</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T01:05:20.697</td>\n",
       "      <td>6208.945257</td>\n",
       "      <td>-9076.852425</td>\n",
       "      <td>5548.296900</td>\n",
       "      <td>4.849212</td>\n",
       "      <td>4.338955</td>\n",
       "      <td>-1.869600</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3931</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T01:26:32.542</td>\n",
       "      <td>10768.200284</td>\n",
       "      <td>-2199.706707</td>\n",
       "      <td>2272.014862</td>\n",
       "      <td>1.940505</td>\n",
       "      <td>6.192887</td>\n",
       "      <td>-3.167724</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     id  sat_id                    epoch         x_sim         y_sim  \\\n",
       "0  3927       1  2014-02-01T00:01:45.162 -13366.891347 -14236.753503   \n",
       "1  3928       1  2014-02-01T00:22:57.007  -7370.434039 -14498.771520   \n",
       "2  3929       1  2014-02-01T00:44:08.852   -572.068654 -13065.289498   \n",
       "3  3930       1  2014-02-01T01:05:20.697   6208.945257  -9076.852425   \n",
       "4  3931       1  2014-02-01T01:26:32.542  10768.200284  -2199.706707   \n",
       "\n",
       "         z_sim    Vx_sim    Vy_sim    Vz_sim  \n",
       "0  6386.774555  4.333815 -0.692764  0.810774  \n",
       "1  7130.411325  5.077413  0.360609  0.313402  \n",
       "2  7033.794876  5.519106  2.012830 -0.539412  \n",
       "3  5548.296900  4.849212  4.338955 -1.869600  \n",
       "4  2272.014862  1.940505  6.192887 -3.167724  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test.head()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Adding some time features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "train['day'] = pd.to_datetime(train.epoch).dt.day\n",
    "test['day'] = pd.to_datetime(test.epoch).dt.day\n",
    "train['hour'] = pd.to_datetime(train.epoch).dt.hour\n",
    "test['hour'] = pd.to_datetime(test.epoch).dt.hour\n",
    "train = train.sort_values(by='day')\n",
    "train.head()\n",
    "\n",
    "sats_to_predict = set(test[\"sat_id\"].unique())\n",
    "\n",
    "categ = ['day', 'hour']\n",
    "other = ['x_sim', 'y_sim', 'z_sim', 'Vx_sim', 'Vy_sim', 'Vz_sim']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.compose import ColumnTransformer\n",
    "\n",
    "column_transformer = ColumnTransformer([\n",
    "    ('ohe', OneHotEncoder(handle_unknown=\"ignore\"), categ),\n",
    "    ('scaling', 'passthrough', other)\n",
    "])\n",
    "\n",
    "column_transformer.fit(train)\n",
    "train_t = column_transformer.transform(train)\n",
    "test_t = column_transformer.transform(test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    plt.hist(train.x, bins=30)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "gamma_km = 398603\n",
    "\n",
    "def transform_row(row, func):\n",
    "    return func(row.values)\n",
    "\n",
    "def transform_dataset(data, func=lambda data: ctk(data, gamma_km), \n",
    "                      columns_from=(\"x\", \"y\", \"z\", \"Vx\", \"Vy\", \"Vz\"), \n",
    "                      columns_to=(\"a\", \"e\", \"inclination\", \"longitude\", \"argument\", \"anomaly\")):\n",
    "    new_data = data.loc[:, list(columns_from)].apply(transform_row, axis=1, result_type=\"expand\", args=(func, ))\n",
    "    new_data.columns = columns_to\n",
    "    return data.drop(list(columns_from), axis=1).join(new_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_quaternion = transform_dataset(train, func=lambda data: ctq(data, gamma_km),\n",
    "                                     columns_from=(\"x\", \"y\", \"z\", \"Vx\", \"Vy\", \"Vz\"),\n",
    "                                     columns_to=((\"a\", \"e\", \"anomaly\", \"q1\", \"q2\", \"q3\", \"q4\")))\n",
    "train_quaternion = transform_dataset(train_quaternion, func=lambda data: ctq(data, gamma_km), \n",
    "                                     columns_from=(\"x_sim\", \"y_sim\", \"z_sim\", \"Vx_sim\", \"Vy_sim\", \"Vz_sim\"),\n",
    "                                     columns_to=(\"a_sim\", \"e_sim\", \"anomaly_sim\", \"q1_sim\", \"q2_sim\", \"q3_sim\", \"q4_sim\"))\n",
    "\n",
    "test_quaternion = transform_dataset(test, func=lambda data: ctq(data, gamma_km), \n",
    "                                    columns_from=(\"x_sim\", \"y_sim\", \"z_sim\", \"Vx_sim\", \"Vy_sim\", \"Vz_sim\"),\n",
    "                                    columns_to=(\"a_sim\", \"e_sim\", \"anomaly_sim\", \"q1_sim\", \"q2_sim\", \"q3_sim\", \"q4_sim\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "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>sat_id</th>\n",
       "      <th>day</th>\n",
       "      <th>a</th>\n",
       "      <th>e</th>\n",
       "      <th>anomaly</th>\n",
       "      <th>q1</th>\n",
       "      <th>q2</th>\n",
       "      <th>q3</th>\n",
       "      <th>q4</th>\n",
       "      <th>a_sim</th>\n",
       "      <th>e_sim</th>\n",
       "      <th>anomaly_sim</th>\n",
       "      <th>q1_sim</th>\n",
       "      <th>q2_sim</th>\n",
       "      <th>q3_sim</th>\n",
       "      <th>q4_sim</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>35750.114408</td>\n",
       "      <td>0.270492</td>\n",
       "      <td>8.560065e-08</td>\n",
       "      <td>-0.593575</td>\n",
       "      <td>0.389910</td>\n",
       "      <td>0.386518</td>\n",
       "      <td>-0.588424</td>\n",
       "      <td>35748.665023</td>\n",
       "      <td>0.270000</td>\n",
       "      <td>0.000784</td>\n",
       "      <td>-0.593442</td>\n",
       "      <td>0.390131</td>\n",
       "      <td>0.386726</td>\n",
       "      <td>-0.588275</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87682</th>\n",
       "      <td>83</td>\n",
       "      <td>1</td>\n",
       "      <td>49685.701978</td>\n",
       "      <td>0.171211</td>\n",
       "      <td>3.716588e+00</td>\n",
       "      <td>-0.449788</td>\n",
       "      <td>0.838916</td>\n",
       "      <td>-0.270081</td>\n",
       "      <td>0.144799</td>\n",
       "      <td>49685.522688</td>\n",
       "      <td>0.170323</td>\n",
       "      <td>3.716723</td>\n",
       "      <td>-0.449794</td>\n",
       "      <td>0.839090</td>\n",
       "      <td>-0.269665</td>\n",
       "      <td>0.144548</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87683</th>\n",
       "      <td>83</td>\n",
       "      <td>1</td>\n",
       "      <td>49684.872139</td>\n",
       "      <td>0.171229</td>\n",
       "      <td>3.922217e+00</td>\n",
       "      <td>-0.449791</td>\n",
       "      <td>0.838916</td>\n",
       "      <td>-0.270077</td>\n",
       "      <td>0.144798</td>\n",
       "      <td>49685.739285</td>\n",
       "      <td>0.170321</td>\n",
       "      <td>3.922636</td>\n",
       "      <td>-0.449794</td>\n",
       "      <td>0.839092</td>\n",
       "      <td>-0.269661</td>\n",
       "      <td>0.144545</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87684</th>\n",
       "      <td>83</td>\n",
       "      <td>1</td>\n",
       "      <td>49684.811687</td>\n",
       "      <td>0.171240</td>\n",
       "      <td>4.140187e+00</td>\n",
       "      <td>-0.449793</td>\n",
       "      <td>0.838916</td>\n",
       "      <td>-0.270074</td>\n",
       "      <td>0.144797</td>\n",
       "      <td>49686.055100</td>\n",
       "      <td>0.170321</td>\n",
       "      <td>4.140800</td>\n",
       "      <td>-0.449794</td>\n",
       "      <td>0.839093</td>\n",
       "      <td>-0.269657</td>\n",
       "      <td>0.144543</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>87685</th>\n",
       "      <td>83</td>\n",
       "      <td>1</td>\n",
       "      <td>49685.602703</td>\n",
       "      <td>0.171246</td>\n",
       "      <td>4.374387e+00</td>\n",
       "      <td>-0.449794</td>\n",
       "      <td>0.838916</td>\n",
       "      <td>-0.270073</td>\n",
       "      <td>0.144797</td>\n",
       "      <td>49686.443023</td>\n",
       "      <td>0.170324</td>\n",
       "      <td>4.375054</td>\n",
       "      <td>-0.449794</td>\n",
       "      <td>0.839094</td>\n",
       "      <td>-0.269655</td>\n",
       "      <td>0.144542</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       sat_id  day             a         e       anomaly        q1        q2  \\\n",
       "0           0    1  35750.114408  0.270492  8.560065e-08 -0.593575  0.389910   \n",
       "87682      83    1  49685.701978  0.171211  3.716588e+00 -0.449788  0.838916   \n",
       "87683      83    1  49684.872139  0.171229  3.922217e+00 -0.449791  0.838916   \n",
       "87684      83    1  49684.811687  0.171240  4.140187e+00 -0.449793  0.838916   \n",
       "87685      83    1  49685.602703  0.171246  4.374387e+00 -0.449794  0.838916   \n",
       "\n",
       "             q3        q4         a_sim     e_sim  anomaly_sim    q1_sim  \\\n",
       "0      0.386518 -0.588424  35748.665023  0.270000     0.000784 -0.593442   \n",
       "87682 -0.270081  0.144799  49685.522688  0.170323     3.716723 -0.449794   \n",
       "87683 -0.270077  0.144798  49685.739285  0.170321     3.922636 -0.449794   \n",
       "87684 -0.270074  0.144797  49686.055100  0.170321     4.140800 -0.449794   \n",
       "87685 -0.270073  0.144797  49686.443023  0.170324     4.375054 -0.449794   \n",
       "\n",
       "         q2_sim    q3_sim    q4_sim  \n",
       "0      0.390131  0.386726 -0.588275  \n",
       "87682  0.839090 -0.269665  0.144548  \n",
       "87683  0.839092 -0.269661  0.144545  \n",
       "87684  0.839093 -0.269657  0.144543  \n",
       "87685  0.839094 -0.269655  0.144542  "
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_quaternion.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "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>sat_id</th>\n",
       "      <th>day</th>\n",
       "      <th>a_sim</th>\n",
       "      <th>e_sim</th>\n",
       "      <th>anomaly_sim</th>\n",
       "      <th>q1_sim</th>\n",
       "      <th>q2_sim</th>\n",
       "      <th>q3_sim</th>\n",
       "      <th>q4_sim</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21111.035679</td>\n",
       "      <td>0.468304</td>\n",
       "      <td>4.276208</td>\n",
       "      <td>-0.964330</td>\n",
       "      <td>0.248479</td>\n",
       "      <td>0.022769</td>\n",
       "      <td>-0.088358</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21110.224464</td>\n",
       "      <td>0.468324</td>\n",
       "      <td>4.558021</td>\n",
       "      <td>-0.964336</td>\n",
       "      <td>0.248466</td>\n",
       "      <td>0.022759</td>\n",
       "      <td>-0.088328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21108.805096</td>\n",
       "      <td>0.468320</td>\n",
       "      <td>4.948948</td>\n",
       "      <td>-0.964345</td>\n",
       "      <td>0.248455</td>\n",
       "      <td>0.022744</td>\n",
       "      <td>-0.088273</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21108.010434</td>\n",
       "      <td>0.468310</td>\n",
       "      <td>5.518140</td>\n",
       "      <td>-0.964351</td>\n",
       "      <td>0.248462</td>\n",
       "      <td>0.022723</td>\n",
       "      <td>-0.088191</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>21110.673154</td>\n",
       "      <td>0.468376</td>\n",
       "      <td>0.002943</td>\n",
       "      <td>-0.964345</td>\n",
       "      <td>0.248497</td>\n",
       "      <td>0.022718</td>\n",
       "      <td>-0.088158</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   sat_id  day         a_sim     e_sim  anomaly_sim    q1_sim    q2_sim  \\\n",
       "0       1    1  21111.035679  0.468304     4.276208 -0.964330  0.248479   \n",
       "1       1    1  21110.224464  0.468324     4.558021 -0.964336  0.248466   \n",
       "2       1    1  21108.805096  0.468320     4.948948 -0.964345  0.248455   \n",
       "3       1    1  21108.010434  0.468310     5.518140 -0.964351  0.248462   \n",
       "4       1    1  21110.673154  0.468376     0.002943 -0.964345  0.248497   \n",
       "\n",
       "     q3_sim    q4_sim  \n",
       "0  0.022769 -0.088358  \n",
       "1  0.022759 -0.088328  \n",
       "2  0.022744 -0.088273  \n",
       "3  0.022723 -0.088191  \n",
       "4  0.022718 -0.088158  "
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_quaternion.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n_train: 519930\n",
      "n_val: 129982\n",
      "n_test: 284071\n"
     ]
    }
   ],
   "source": [
    "# target variables columns \n",
    "target_cols = ['x', 'y', 'z', 'Vx', 'Vy', 'Vz']\n",
    "#target_cols = ['a', 'e', 'anomaly', 'q1', 'q2', 'q3', 'q4',]\n",
    "# features columns \n",
    "feature_cols = [c for c in train.columns if c not in target_cols and c != 'epoch' and c != 'sat_id']\n",
    "# number of rows for training, validation and prediction\n",
    "n_train = train.shape[0]\n",
    "n_val = int(0.2 * n_train)\n",
    "n_train = n_train - n_val\n",
    "n_test = test.shape[0]\n",
    "\n",
    "print(f\"n_train: {n_train}\\nn_val: {n_val}\\nn_test: {n_test}\")\n",
    "#assert n_train + n_val == train.shape[0]\n",
    "\n",
    "# train, validation, test arrays\n",
    "X_train = train[feature_cols].values\n",
    "y_train = dict(zip(target_cols, [train[l].values for l in target_cols]))\n",
    "X_val = train[feature_cols].values[n_train:n_train+n_val]\n",
    "y_val = dict(zip(target_cols, [train[l].values[n_train:n_train+n_val] for l in target_cols]))\n",
    "X_test = test[feature_cols].values\n",
    "\n",
    "#assert X_train.shape[0] == y_train[target_cols[0]].shape[0] == n_train\n",
    "assert X_val.shape[0] == y_val[target_cols[0]].shape[0] == n_val\n",
    "assert X_test.shape[0] == n_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predict 6 coordinates one by one (worked slightly better)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_one(train_quaternion, test_quaternion, degree=1):\n",
    "    #train_sat = train_quaternion[train_quaternion.sat_id == sat_id]\n",
    "    #test_sat = test_quaternion[test_quaternion.sat_id == sat_id]\n",
    "    result = pd.DataFrame(columns=[\"id\", 'x', 'y', 'z', 'Vx', 'Vy', 'Vz'])\n",
    "    \n",
    "    result[\"id\"] = test[\"id\"]\n",
    "    \n",
    "    train_features = ['hour', 'day', 'x_sim', 'y_sim', 'z_sim', 'Vx_sim', 'Vy_sim', 'Vz_sim']\n",
    "    test_features = ['x_sim', 'y_sim', 'z_sim', 'Vx_sim', 'Vy_sim', 'Vz_sim']\n",
    "    for coordinate in ['x', 'y', 'z', 'Vx', 'Vy', 'Vz']:\n",
    "        train_diff = train[coordinate] - train[coordinate + \"_sim\"]\n",
    "        train_diff = np.array(train_diff).reshape(-1, 1)\n",
    "\n",
    "        model = Ridge()\n",
    "        train_t = column_transformer.transform(train[train_features])\n",
    "        model.fit(train_t, train_diff)\n",
    "\n",
    "        test_t = column_transformer.transform(test[train_features])\n",
    "        #print(test[coordinate + \"_sim\"].shape, model.predict(test_t).shape)\n",
    "        result[coordinate] = test[coordinate + \"_sim\"] + model.predict(test_t).reshape(-1, )\n",
    "        \n",
    "    #train_diff = phase_regression(train_sat[\"anomaly\"].to_numpy()) - phase_regression(train_sat[\"anomaly_sim\"].to_numpy())\n",
    "    #model = LinearRegression().fit(train_features, train_diff)\n",
    "    #result[\"anomaly\"] = phase_degression(test_sat[\"anomaly_sim\"].to_numpy() + model.predict(test_features))\n",
    "    \n",
    "    #if np.any(result[\"e\"].to_numpy() >= 1.) or np.any(result[\"e\"].to_numpy() < 0.):\n",
    "        #result[\"e\"] = test_sat[\"e_sim\"]\n",
    "\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "results = []\n",
    "\n",
    "#for sat_id in tqdm(sats_to_predict):\n",
    "#    results.append(predict_one(train, test, sat_id))\n",
    "    \n",
    "result = predict_one(train, test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "result.index = result.id\n",
    "result.drop('id', axis=1, inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "result.to_csv('predictions.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>id</th>\n",
       "      <th>x</th>\n",
       "      <th>y</th>\n",
       "      <th>z</th>\n",
       "      <th>Vx</th>\n",
       "      <th>Vy</th>\n",
       "      <th>Vz</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3927</td>\n",
       "      <td>-12595.531033</td>\n",
       "      <td>-13860.943848</td>\n",
       "      <td>5926.795414</td>\n",
       "      <td>4.286915</td>\n",
       "      <td>-0.731256</td>\n",
       "      <td>0.841993</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3928</td>\n",
       "      <td>-6837.721178</td>\n",
       "      <td>-14050.884151</td>\n",
       "      <td>6670.825913</td>\n",
       "      <td>5.034048</td>\n",
       "      <td>0.323971</td>\n",
       "      <td>0.354062</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3929</td>\n",
       "      <td>-323.744368</td>\n",
       "      <td>-12593.311081</td>\n",
       "      <td>6634.492157</td>\n",
       "      <td>5.482309</td>\n",
       "      <td>1.977449</td>\n",
       "      <td>-0.492386</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3930</td>\n",
       "      <td>6156.493989</td>\n",
       "      <td>-8662.160840</td>\n",
       "      <td>5299.187251</td>\n",
       "      <td>4.826517</td>\n",
       "      <td>4.303513</td>\n",
       "      <td>-1.823344</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3931</td>\n",
       "      <td>10488.809397</td>\n",
       "      <td>-1946.679157</td>\n",
       "      <td>2265.186777</td>\n",
       "      <td>1.944092</td>\n",
       "      <td>6.161347</td>\n",
       "      <td>-3.136209</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284066</th>\n",
       "      <td>1231060</td>\n",
       "      <td>28604.333705</td>\n",
       "      <td>-81641.729674</td>\n",
       "      <td>4228.493051</td>\n",
       "      <td>-1.627582</td>\n",
       "      <td>-0.077548</td>\n",
       "      <td>0.212984</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284067</th>\n",
       "      <td>1231061</td>\n",
       "      <td>16051.655589</td>\n",
       "      <td>-80655.176090</td>\n",
       "      <td>4969.271686</td>\n",
       "      <td>-1.741344</td>\n",
       "      <td>0.263617</td>\n",
       "      <td>0.184210</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284068</th>\n",
       "      <td>1231062</td>\n",
       "      <td>2967.629684</td>\n",
       "      <td>-76903.353861</td>\n",
       "      <td>5539.422507</td>\n",
       "      <td>-1.807036</td>\n",
       "      <td>0.662737</td>\n",
       "      <td>0.144367</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284069</th>\n",
       "      <td>1231063</td>\n",
       "      <td>-10212.556021</td>\n",
       "      <td>-69942.255685</td>\n",
       "      <td>5877.921092</td>\n",
       "      <td>-1.798321</td>\n",
       "      <td>1.134456</td>\n",
       "      <td>0.090226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284070</th>\n",
       "      <td>1231064</td>\n",
       "      <td>-22799.676541</td>\n",
       "      <td>-59227.315392</td>\n",
       "      <td>5899.810433</td>\n",
       "      <td>-1.666065</td>\n",
       "      <td>1.693754</td>\n",
       "      <td>0.017062</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>284071 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             id             x             y            z        Vx        Vy  \\\n",
       "0          3927 -12595.531033 -13860.943848  5926.795414  4.286915 -0.731256   \n",
       "1          3928  -6837.721178 -14050.884151  6670.825913  5.034048  0.323971   \n",
       "2          3929   -323.744368 -12593.311081  6634.492157  5.482309  1.977449   \n",
       "3          3930   6156.493989  -8662.160840  5299.187251  4.826517  4.303513   \n",
       "4          3931  10488.809397  -1946.679157  2265.186777  1.944092  6.161347   \n",
       "...         ...           ...           ...          ...       ...       ...   \n",
       "284066  1231060  28604.333705 -81641.729674  4228.493051 -1.627582 -0.077548   \n",
       "284067  1231061  16051.655589 -80655.176090  4969.271686 -1.741344  0.263617   \n",
       "284068  1231062   2967.629684 -76903.353861  5539.422507 -1.807036  0.662737   \n",
       "284069  1231063 -10212.556021 -69942.255685  5877.921092 -1.798321  1.134456   \n",
       "284070  1231064 -22799.676541 -59227.315392  5899.810433 -1.666065  1.693754   \n",
       "\n",
       "              Vz  \n",
       "0       0.841993  \n",
       "1       0.354062  \n",
       "2      -0.492386  \n",
       "3      -1.823344  \n",
       "4      -3.136209  \n",
       "...          ...  \n",
       "284066  0.212984  \n",
       "284067  0.184210  \n",
       "284068  0.144367  \n",
       "284069  0.090226  \n",
       "284070  0.017062  \n",
       "\n",
       "[284071 rows x 7 columns]"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "result_quaternion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "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>id</th>\n",
       "      <th>sat_id</th>\n",
       "      <th>epoch</th>\n",
       "      <th>x_sim</th>\n",
       "      <th>y_sim</th>\n",
       "      <th>z_sim</th>\n",
       "      <th>Vx_sim</th>\n",
       "      <th>Vy_sim</th>\n",
       "      <th>Vz_sim</th>\n",
       "      <th>day</th>\n",
       "      <th>hour</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>3927</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:01:45.162</td>\n",
       "      <td>-13366.891347</td>\n",
       "      <td>-14236.753503</td>\n",
       "      <td>6386.774555</td>\n",
       "      <td>4.333815</td>\n",
       "      <td>-0.692764</td>\n",
       "      <td>0.810774</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>3928</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:22:57.007</td>\n",
       "      <td>-7370.434039</td>\n",
       "      <td>-14498.771520</td>\n",
       "      <td>7130.411325</td>\n",
       "      <td>5.077413</td>\n",
       "      <td>0.360609</td>\n",
       "      <td>0.313402</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3929</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T00:44:08.852</td>\n",
       "      <td>-572.068654</td>\n",
       "      <td>-13065.289498</td>\n",
       "      <td>7033.794876</td>\n",
       "      <td>5.519106</td>\n",
       "      <td>2.012830</td>\n",
       "      <td>-0.539412</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3930</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T01:05:20.697</td>\n",
       "      <td>6208.945257</td>\n",
       "      <td>-9076.852425</td>\n",
       "      <td>5548.296900</td>\n",
       "      <td>4.849212</td>\n",
       "      <td>4.338955</td>\n",
       "      <td>-1.869600</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3931</td>\n",
       "      <td>1</td>\n",
       "      <td>2014-02-01T01:26:32.542</td>\n",
       "      <td>10768.200284</td>\n",
       "      <td>-2199.706707</td>\n",
       "      <td>2272.014862</td>\n",
       "      <td>1.940505</td>\n",
       "      <td>6.192887</td>\n",
       "      <td>-3.167724</td>\n",
       "      <td>1</td>\n",
       "      <td>1</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",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284066</th>\n",
       "      <td>1231060</td>\n",
       "      <td>597</td>\n",
       "      <td>2014-02-28T05:19:01.386</td>\n",
       "      <td>28595.031634</td>\n",
       "      <td>-85458.623976</td>\n",
       "      <td>5021.767767</td>\n",
       "      <td>-1.721131</td>\n",
       "      <td>-0.011611</td>\n",
       "      <td>0.122357</td>\n",
       "      <td>28</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284067</th>\n",
       "      <td>1231061</td>\n",
       "      <td>597</td>\n",
       "      <td>2014-02-28T07:21:46.454</td>\n",
       "      <td>15547.173728</td>\n",
       "      <td>-84233.509948</td>\n",
       "      <td>5840.616702</td>\n",
       "      <td>-1.815323</td>\n",
       "      <td>0.353445</td>\n",
       "      <td>0.098927</td>\n",
       "      <td>28</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284068</th>\n",
       "      <td>1231062</td>\n",
       "      <td>597</td>\n",
       "      <td>2014-02-28T09:24:31.522</td>\n",
       "      <td>1981.107111</td>\n",
       "      <td>-80123.860051</td>\n",
       "      <td>6458.394351</td>\n",
       "      <td>-1.858963</td>\n",
       "      <td>0.773846</td>\n",
       "      <td>0.067379</td>\n",
       "      <td>28</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284069</th>\n",
       "      <td>1231063</td>\n",
       "      <td>597</td>\n",
       "      <td>2014-02-28T11:27:16.590</td>\n",
       "      <td>-11644.801379</td>\n",
       "      <td>-72674.504171</td>\n",
       "      <td>6805.361999</td>\n",
       "      <td>-1.826030</td>\n",
       "      <td>1.262930</td>\n",
       "      <td>0.024782</td>\n",
       "      <td>28</td>\n",
       "      <td>11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>284070</th>\n",
       "      <td>1231064</td>\n",
       "      <td>597</td>\n",
       "      <td>2014-02-28T13:30:01.658</td>\n",
       "      <td>-24606.415254</td>\n",
       "      <td>-61329.599053</td>\n",
       "      <td>6784.763455</td>\n",
       "      <td>-1.667849</td>\n",
       "      <td>1.833746</td>\n",
       "      <td>-0.033371</td>\n",
       "      <td>28</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>284071 rows × 11 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "             id  sat_id                    epoch         x_sim         y_sim  \\\n",
       "0          3927       1  2014-02-01T00:01:45.162 -13366.891347 -14236.753503   \n",
       "1          3928       1  2014-02-01T00:22:57.007  -7370.434039 -14498.771520   \n",
       "2          3929       1  2014-02-01T00:44:08.852   -572.068654 -13065.289498   \n",
       "3          3930       1  2014-02-01T01:05:20.697   6208.945257  -9076.852425   \n",
       "4          3931       1  2014-02-01T01:26:32.542  10768.200284  -2199.706707   \n",
       "...         ...     ...                      ...           ...           ...   \n",
       "284066  1231060     597  2014-02-28T05:19:01.386  28595.031634 -85458.623976   \n",
       "284067  1231061     597  2014-02-28T07:21:46.454  15547.173728 -84233.509948   \n",
       "284068  1231062     597  2014-02-28T09:24:31.522   1981.107111 -80123.860051   \n",
       "284069  1231063     597  2014-02-28T11:27:16.590 -11644.801379 -72674.504171   \n",
       "284070  1231064     597  2014-02-28T13:30:01.658 -24606.415254 -61329.599053   \n",
       "\n",
       "              z_sim    Vx_sim    Vy_sim    Vz_sim  day  hour  \n",
       "0       6386.774555  4.333815 -0.692764  0.810774    1     0  \n",
       "1       7130.411325  5.077413  0.360609  0.313402    1     0  \n",
       "2       7033.794876  5.519106  2.012830 -0.539412    1     0  \n",
       "3       5548.296900  4.849212  4.338955 -1.869600    1     1  \n",
       "4       2272.014862  1.940505  6.192887 -3.167724    1     1  \n",
       "...             ...       ...       ...       ...  ...   ...  \n",
       "284066  5021.767767 -1.721131 -0.011611  0.122357   28     5  \n",
       "284067  5840.616702 -1.815323  0.353445  0.098927   28     7  \n",
       "284068  6458.394351 -1.858963  0.773846  0.067379   28     9  \n",
       "284069  6805.361999 -1.826030  1.262930  0.024782   28    11  \n",
       "284070  6784.763455 -1.667849  1.833746 -0.033371   28    13  \n",
       "\n",
       "[284071 rows x 11 columns]"
      ]
     },
     "execution_count": 59,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Predict 6 coordinates at one time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "x\n",
      "MAPE train:  153.7868452280808\n",
      "MAPE val:  196.61063647438687 \n",
      "\n",
      "y\n",
      "MAPE train:  317.18745607326355\n",
      "MAPE val:  267.42317790057643 \n",
      "\n",
      "z\n",
      "MAPE train:  321.97972152877765\n",
      "MAPE val:  350.56668013519936 \n",
      "\n",
      "Vx\n",
      "MAPE train:  191.25675329409856\n",
      "MAPE val:  315.7228145682371 \n",
      "\n",
      "Vy\n",
      "MAPE train:  134.42911944517954\n",
      "MAPE val:  192.86740383289666 \n",
      "\n",
      "Vz\n",
      "MAPE train:  1035.6067536033372\n",
      "MAPE val:  463.76238044126967 \n",
      "\n",
      "------\n",
      "mean MAPE train:  359.0411081954562\n",
      "mean MAPE val:  297.82551555876097\n"
     ]
    }
   ],
   "source": [
    "# dictionary: {variable: LR predictions}\n",
    "y_test = {}\n",
    "\n",
    "_mape_train_list = []\n",
    "_mape_val_list = []\n",
    "for k, v in y_train.items():\n",
    "    model = LinearRegression().fit(X_train, v)\n",
    "    mape_train = np.mean(np.abs((model.predict(X_train) - v) / v)) * 100\n",
    "    _mape_train_list.append(mape_train)\n",
    "    mape_val = np.mean(np.abs((model.predict(X_val) - y_val[k]) / y_val[k])) * 100\n",
    "    _mape_val_list.append(mape_val)\n",
    "    model = LinearRegression().fit(\n",
    "        np.vstack([X_train, X_val]), np.hstack([v, y_val[k]]))\n",
    "    y_test[k] = model.predict(X_test)\n",
    "    print(k)\n",
    "    print(\"MAPE train: \", mape_train)\n",
    "    print(\"MAPE val: \", mape_val, \"\\n\")\n",
    "\n",
    "print(\"------\")\n",
    "print(\"mean MAPE train: \", np.mean(_mape_train_list))\n",
    "print(\"mean MAPE val: \", np.mean(_mape_val_list))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "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>x</th>\n",
       "      <th>y</th>\n",
       "      <th>z</th>\n",
       "      <th>Vx</th>\n",
       "      <th>Vy</th>\n",
       "      <th>Vz</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>3927</th>\n",
       "      <td>-13670.775275</td>\n",
       "      <td>-14378.827605</td>\n",
       "      <td>5138.391144</td>\n",
       "      <td>3.558354</td>\n",
       "      <td>-0.545464</td>\n",
       "      <td>0.874702</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3928</th>\n",
       "      <td>-8071.832174</td>\n",
       "      <td>-14848.017677</td>\n",
       "      <td>6147.384454</td>\n",
       "      <td>4.215326</td>\n",
       "      <td>0.330379</td>\n",
       "      <td>0.498611</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3929</th>\n",
       "      <td>-1669.322349</td>\n",
       "      <td>-13742.150723</td>\n",
       "      <td>6601.844753</td>\n",
       "      <td>4.640011</td>\n",
       "      <td>1.694490</td>\n",
       "      <td>-0.187929</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3930</th>\n",
       "      <td>4888.221187</td>\n",
       "      <td>-10138.980516</td>\n",
       "      <td>6083.159497</td>\n",
       "      <td>4.160835</td>\n",
       "      <td>3.598294</td>\n",
       "      <td>-1.319466</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3931</th>\n",
       "      <td>9668.065944</td>\n",
       "      <td>-3375.682862</td>\n",
       "      <td>3931.941334</td>\n",
       "      <td>1.808036</td>\n",
       "      <td>5.090177</td>\n",
       "      <td>-2.521150</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 x             y            z        Vx        Vy        Vz\n",
       "id                                                                         \n",
       "3927 -13670.775275 -14378.827605  5138.391144  3.558354 -0.545464  0.874702\n",
       "3928  -8071.832174 -14848.017677  6147.384454  4.215326  0.330379  0.498611\n",
       "3929  -1669.322349 -13742.150723  6601.844753  4.640011  1.694490 -0.187929\n",
       "3930   4888.221187 -10138.980516  6083.159497  4.160835  3.598294 -1.319466\n",
       "3931   9668.065944  -3375.682862  3931.941334  1.808036  5.090177 -2.521150"
      ]
     },
     "execution_count": 104,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "res = pd.DataFrame()\n",
    "\n",
    "res['x'] = y_test['x']\n",
    "res['y'] = y_test['y']\n",
    "res['z'] = y_test['z']\n",
    "res['Vx'] = y_test['Vx']\n",
    "res['Vy'] = y_test['Vy']\n",
    "res['Vz'] = y_test['Vz']\n",
    "res.index = test.id\n",
    "predictions = pd.DataFrame(res, index=test.id[-n_test:])\n",
    "\n",
    "res.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "result_cartesian = transform_dataset(res, func=lambda data: qtc(data, gamma_km), \n",
    "                                     columns_from=(\"a\", \"e\", \"anomaly\", \"q1\", \"q2\", \"q3\", \"q4\"),\n",
    "                                     columns_to=(\"x\", \"y\", \"z\", \"Vx\", \"Vy\", \"Vz\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "# predictions\n",
    "predictions = pd.DataFrame(result_cartesian, index=test.index[-n_test:])\n",
    "predictions.to_csv('predictions.csv')"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}