Models.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "8f33bcba",
   "metadata": {
    "id": "8f33bcba"
   },
   "outputs": [],
   "source": [
    "#Basic imports\n",
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import warnings\n",
    "import time\n",
    "import joblib\n",
    "\n",
    "import helpers as helpers"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46a620d1",
   "metadata": {
    "id": "46a620d1"
   },
   "source": [
    "# Data Processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "378a854b",
   "metadata": {
    "id": "378a854b"
   },
   "outputs": [],
   "source": [
    "# Import data\n",
    "data1 = pd.read_excel('data.xlsx', 'Timepoint 1', header=1)\n",
    "data2 = pd.read_excel('data2.xlsx', 'Sheet1', header=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "5ef5a083",
   "metadata": {
    "id": "5ef5a083",
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "data2['Date of birth'] = pd.to_datetime(data2['Date of birth'].astype(str), format='%d/%m/%Y')\n",
    "data2['Date of MRI scan'] = pd.to_datetime(data2['Date of MRI scan'].astype(str), format='%Y-%m-%d', errors='ignore')\n",
    "data2['Age'] = (data2['Date of MRI scan'] - data2['Date of birth'])/ np.timedelta64(1, 'Y')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7fc0feef",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset1 : (133 samples, 335 features)\n",
      "Dataset2 : (232 samples, 175 features)\n"
     ]
    }
   ],
   "source": [
    "sns.kdeplot(\n",
    "    data1,\n",
    "    x='Age',\n",
    "    label='Dataset 1'\n",
    ").set_title(\"Age distribution of datasets\")\n",
    "\n",
    "sns.kdeplot(\n",
    "    data2,\n",
    "    x='Age',\n",
    "    label='Dataset 2'\n",
    ")\n",
    "\n",
    "plt.legend()\n",
    "plt.show()\n",
    "\n",
    "print(f'Dataset1 : ({data1.shape[0]} samples, {data1.shape[1]} features)')\n",
    "print(f'Dataset2 : ({data2.shape[0]} samples, {data2.shape[1]} features)')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b4eb3e7b",
   "metadata": {},
   "source": [
    "## Format dataset 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1ed420e3",
   "metadata": {
    "id": "1ed420e3"
   },
   "outputs": [],
   "source": [
    "data_train1, data_test1, y_train1, y_test1 = train_test_split(data1.drop(['Age'], axis=1), data1['Age'], test_size=0.3, random_state=2)\n",
    "x_train1, del_col1, mean_train1, std_train1 = helpers.clean_data(1, data_train1)\n",
    "x_test1, _, _, _ = helpers.clean_data(1, data_test1, del_col1, mean_train1, std_train1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "39fd4c9f",
   "metadata": {
    "id": "39fd4c9f"
   },
   "outputs": [],
   "source": [
    "x_train1 = x_train1.reset_index(drop=True)\n",
    "x_test1 = x_test1.reset_index(drop=True)\n",
    "y_train1 = pd.DataFrame(data=y_train1, columns=['Age']).reset_index(drop=True)\n",
    "y_test1 = pd.DataFrame(data=y_test1, columns=['Age']).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67af90bd",
   "metadata": {},
   "source": [
    "## Format dataset 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7c464f08",
   "metadata": {},
   "outputs": [],
   "source": [
    "data_train2, data_test2, y_train2, y_test2 = train_test_split(data2.drop(['Age'], axis=1), data2['Age'], test_size=0.3, random_state=2)\n",
    "x_train2, del_col2, mean_train2, std_train2 = helpers.clean_data(2, data_train2)\n",
    "x_test2, _, _, _ = helpers.clean_data(2, data_test2, del_col2, mean_train2, std_train2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "108b6e1d",
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train2 = x_train2.reset_index(drop=True)\n",
    "x_test2 = x_test2.reset_index(drop=True)\n",
    "y_train2 = pd.DataFrame(data=y_train2, columns=['Age']).reset_index(drop=True)\n",
    "y_test2 = pd.DataFrame(data=y_test2, columns=['Age']).reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8806fa41",
   "metadata": {},
   "source": [
    "## Select dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e4cba3d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "i_data = 1\n",
    "\n",
    "if (i_data == 1):\n",
    "    x_train = x_train1\n",
    "    x_test = x_test1\n",
    "    y_train = y_train1\n",
    "    y_test = y_test1\n",
    "elif (i_data == 2):\n",
    "    x_train = x_train2\n",
    "    x_test = x_test2\n",
    "    y_train = y_train2\n",
    "    y_test = y_test2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "89e290d0",
   "metadata": {
    "id": "89e290d0"
   },
   "source": [
    "### Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "002589ee",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import svm\n",
    "from sklearn.linear_model import ElasticNet\n",
    "from sklearn.ensemble import AdaBoostRegressor\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.neighbors import KNeighborsRegressor\n",
    "\n",
    "# Create grids for search parameters\n",
    "param_grid_svm = [\n",
    "    {'clf__C': np.logspace(-3, 3, 10), 'clf__kernel': ['linear']},\n",
    "    {'clf__C': np.logspace(-3, 3, 10), 'clf__gamma': np.logspace(-10, 0, 20), 'clf__kernel': ['rbf']},\n",
    "]\n",
    "\n",
    "param_grid_elastic = {\n",
    "    'clf__alpha': np.logspace(-10,7,25),\n",
    "    'clf__l1_ratio': np.append(np.logspace(-10,0,20), 0)\n",
    "}\n",
    "\n",
    "param_grid_adaboost = {\n",
    "    'clf__n_estimators': np.arange(1, 200, 20),\n",
    "    'clf__learning_rate': [1.0, 0.5, 0.1]\n",
    "}\n",
    "\n",
    "param_grid_forest = {\n",
    "    'clf__n_estimators': np.arange(10, 100, 20),\n",
    "    'clf__max_depth': np.arange(1, 20)\n",
    "}\n",
    "\n",
    "param_grid_knn = {\n",
    "    'clf__n_neighbors': np.arange(1, 20)\n",
    "}\n",
    "\n",
    "\n",
    "# Create models grid search\n",
    "model_svr = helpers.model_gs(\n",
    "    svm.SVR(),\n",
    "    param_grid_svm\n",
    ")\n",
    "\n",
    "model_elastic = helpers.model_gs(\n",
    "    ElasticNet(max_iter=5000),\n",
    "    param_grid_elastic\n",
    ")\n",
    "\n",
    "model_adaboost = helpers.model_gs(\n",
    "    AdaBoostRegressor(random_state = 123),\n",
    "    param_grid_adaboost\n",
    ")\n",
    "\n",
    "model_forest = helpers.model_gs(\n",
    "    RandomForestRegressor(),\n",
    "    param_grid_forest\n",
    ")\n",
    "\n",
    "model_knn = helpers.model_gs(\n",
    "    KNeighborsRegressor(),\n",
    "    param_grid_knn\n",
    ")\n",
    "\n",
    "# Create models list\n",
    "models = {\n",
    "    \"SVR\": model_svr,\n",
    "    \"ElasticNet\": model_elastic,\n",
    "    \"AdaBoost\": model_adaboost,\n",
    "    \"RandomForest\": model_forest,\n",
    "    \"KNN\": model_knn\n",
    "}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a122fc28",
   "metadata": {
    "id": "a122fc28"
   },
   "source": [
    "# PLS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "78af5de4",
   "metadata": {
    "id": "78af5de4"
   },
   "outputs": [],
   "source": [
    "# PLS Regression\n",
    "pls_train, pls_test = helpers.PLS_regression(x_train, x_test, y_train, y_test, n_components = 20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10c12033",
   "metadata": {
    "id": "10c12033"
   },
   "source": [
    "### Standardize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "aa167953",
   "metadata": {
    "id": "aa167953"
   },
   "outputs": [],
   "source": [
    "pls_train_scaled, pls_test_scaled = helpers.standardize(pls_train, pls_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed2bea9d",
   "metadata": {},
   "source": [
    "### Filter MI Score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "77655acd",
   "metadata": {
    "id": "77655acd"
   },
   "outputs": [],
   "source": [
    "pls_train_scaled, pls_test_scaled = helpers.filter_mi_scores(pls_train_scaled, pls_test_scaled, y_train, 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83cbee66",
   "metadata": {
    "id": "83cbee66"
   },
   "source": [
    "### Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "0b8c9e79",
   "metadata": {
    "id": "0b8c9e79"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:870: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "pls_train_scaled, pls_test_scaled = helpers.KM_clustering(pls_train_scaled, pls_test_scaled, n_clusters=3)\n",
    "#pls_train_scaled, pls_test_scaled = helpers.GMM_clustering(pls_train_scaled, pls_test_scaled, n_clusters=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "f9a6246e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_pls_global = helpers.train_all(pls_train_scaled, pls_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "4292ec71",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the test scores are non-finite: [-0.45109469 -0.18498202 -0.00572451  0.20451302  0.25446138  0.32697635\n",
      "  0.33797776  0.30207323  0.30672695  0.30057545  0.28614114  0.27263376\n",
      "  0.21964039  0.13759289  0.09888096  0.01393505 -0.05378727 -0.13567855\n",
      "         nan]\n",
      "  warnings.warn(\n",
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the train scores are non-finite: [1.         0.77248375 0.68096804 0.58691028 0.55475449 0.50565638\n",
      " 0.45688171 0.42302798 0.36776405 0.32753201 0.29683506 0.26639379\n",
      " 0.22645064 0.18991134 0.15112557 0.11547554 0.07373773 0.02141678\n",
      "        nan]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n",
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_pls_cluster_0 = helpers.train_cluster(0, pls_train_scaled, pls_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_pls_cluster_1 = helpers.train_cluster(1, pls_train_scaled, pls_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_pls_cluster_2 = helpers.train_cluster(2, pls_train_scaled, pls_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "f36fc8a6",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.875250</td>\n",
       "      <td>0.929125</td>\n",
       "      <td>-0.141497</td>\n",
       "      <td>0.673725</td>\n",
       "      <td>3.099315</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.892625</td>\n",
       "      <td>0.922724</td>\n",
       "      <td>-0.101993</td>\n",
       "      <td>0.741389</td>\n",
       "      <td>3.067188</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.517537</td>\n",
       "      <td>0.931520</td>\n",
       "      <td>0.075040</td>\n",
       "      <td>0.734198</td>\n",
       "      <td>2.782801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.518894</td>\n",
       "      <td>0.944438</td>\n",
       "      <td>-0.012450</td>\n",
       "      <td>0.590323</td>\n",
       "      <td>2.917500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.399964</td>\n",
       "      <td>0.731538</td>\n",
       "      <td>-0.033066</td>\n",
       "      <td>1.346774</td>\n",
       "      <td>2.737500</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.875250  0.929125 -0.141497   0.673725  3.099315\n",
       "ElasticNet      0.892625  0.922724 -0.101993   0.741389  3.067188\n",
       "AdaBoost        0.517537  0.931520  0.075040   0.734198  2.782801\n",
       "RandomForest    0.518894  0.944438 -0.012450   0.590323  2.917500\n",
       "KNN             0.399964  0.731538 -0.033066   1.346774  2.737500"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pls_global"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "e9cfe6c2",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.768469</td>\n",
       "      <td>0.941004</td>\n",
       "      <td>-0.073342</td>\n",
       "      <td>0.435212</td>\n",
       "      <td>2.590544</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.803905</td>\n",
       "      <td>0.948954</td>\n",
       "      <td>-0.095828</td>\n",
       "      <td>0.546679</td>\n",
       "      <td>2.620294</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.139751</td>\n",
       "      <td>0.970117</td>\n",
       "      <td>-0.082903</td>\n",
       "      <td>0.420946</td>\n",
       "      <td>2.327434</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.299485</td>\n",
       "      <td>0.907810</td>\n",
       "      <td>-0.022324</td>\n",
       "      <td>0.671429</td>\n",
       "      <td>2.304000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.337978</td>\n",
       "      <td>0.573431</td>\n",
       "      <td>0.011154</td>\n",
       "      <td>1.454082</td>\n",
       "      <td>2.428571</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.768469  0.941004 -0.073342   0.435212  2.590544\n",
       "ElasticNet      0.803905  0.948954 -0.095828   0.546679  2.620294\n",
       "AdaBoost        0.139751  0.970117 -0.082903   0.420946  2.327434\n",
       "RandomForest    0.299485  0.907810 -0.022324   0.671429  2.304000\n",
       "KNN             0.337978  0.573431  0.011154   1.454082  2.428571"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pls_cluster_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "0c562f64",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.754053</td>\n",
       "      <td>0.940545</td>\n",
       "      <td>-0.220536</td>\n",
       "      <td>0.507179</td>\n",
       "      <td>3.452131</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.817886</td>\n",
       "      <td>0.947409</td>\n",
       "      <td>-0.321073</td>\n",
       "      <td>0.618647</td>\n",
       "      <td>3.609627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.483568</td>\n",
       "      <td>0.984036</td>\n",
       "      <td>-0.070348</td>\n",
       "      <td>0.270562</td>\n",
       "      <td>3.057557</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.403307</td>\n",
       "      <td>0.955687</td>\n",
       "      <td>-0.137468</td>\n",
       "      <td>0.542708</td>\n",
       "      <td>3.176667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.263417</td>\n",
       "      <td>0.616459</td>\n",
       "      <td>-0.117266</td>\n",
       "      <td>1.671875</td>\n",
       "      <td>3.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.754053  0.940545 -0.220536   0.507179  3.452131\n",
       "ElasticNet      0.817886  0.947409 -0.321073   0.618647  3.609627\n",
       "AdaBoost        0.483568  0.984036 -0.070348   0.270562  3.057557\n",
       "RandomForest    0.403307  0.955687 -0.137468   0.542708  3.176667\n",
       "KNN             0.263417  0.616459 -0.117266   1.671875  3.000000"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pls_cluster_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "37bfee9a",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.806907</td>\n",
       "      <td>0.978109</td>\n",
       "      <td>0.290924</td>\n",
       "      <td>0.323294</td>\n",
       "      <td>1.852079</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.827398</td>\n",
       "      <td>0.931224</td>\n",
       "      <td>0.452264</td>\n",
       "      <td>0.737093</td>\n",
       "      <td>1.684856</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.396628</td>\n",
       "      <td>0.977255</td>\n",
       "      <td>-0.769926</td>\n",
       "      <td>0.389048</td>\n",
       "      <td>3.669608</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.466889</td>\n",
       "      <td>0.948632</td>\n",
       "      <td>-0.280721</td>\n",
       "      <td>0.648485</td>\n",
       "      <td>3.168889</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.253650</td>\n",
       "      <td>0.687746</td>\n",
       "      <td>0.144345</td>\n",
       "      <td>1.505051</td>\n",
       "      <td>2.466667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.806907  0.978109  0.290924   0.323294  1.852079\n",
       "ElasticNet      0.827398  0.931224  0.452264   0.737093  1.684856\n",
       "AdaBoost        0.396628  0.977255 -0.769926   0.389048  3.669608\n",
       "RandomForest    0.466889  0.948632 -0.280721   0.648485  3.168889\n",
       "KNN             0.253650  0.687746  0.144345   1.505051  2.466667"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pls_cluster_2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8ab0295f",
   "metadata": {
    "id": "8ab0295f"
   },
   "source": [
    "# PCA"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "0a1c0ac9",
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 404
    },
    "id": "0a1c0ac9",
    "outputId": "9ce73e8b-be62-48cf-e64d-e8ef56b641d4"
   },
   "outputs": [],
   "source": [
    "pc_train, pc_test = helpers.PCA_decomposition(x_train, x_test, y_train, y_test, n_components=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97e87996",
   "metadata": {
    "id": "97e87996"
   },
   "source": [
    "### Standardize"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "917b8131",
   "metadata": {
    "id": "917b8131"
   },
   "outputs": [],
   "source": [
    "pc_train_scaled, pc_test_scaled = helpers.standardize(pc_train, pc_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3eb3bd84",
   "metadata": {
    "id": "3eb3bd84"
   },
   "source": [
    "### Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "f75a8edf",
   "metadata": {
    "id": "f75a8edf"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:870: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "pc_train_scaled, pc_test_scaled = helpers.KM_clustering(pc_train_scaled, pc_test_scaled, n_clusters=3)\n",
    "#pc_train_scaled, pc_test_scaled = helpers.GMM_clustering(pc_train_scaled, pc_test_scaled, n_clusters=3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "946b7253",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 3.919e+02, tolerance: 1.130e-01 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_pca_global = helpers.train_all(pc_train_scaled, pc_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "b832d1ee",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n",
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\metrics\\_regression.py:918: UndefinedMetricWarning: R^2 score is not well-defined with less than two samples.\n",
      "  warnings.warn(msg, UndefinedMetricWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\metrics\\_regression.py:918: UndefinedMetricWarning: R^2 score is not well-defined with less than two samples.\n",
      "  warnings.warn(msg, UndefinedMetricWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\metrics\\_regression.py:918: UndefinedMetricWarning: R^2 score is not well-defined with less than two samples.\n",
      "  warnings.warn(msg, UndefinedMetricWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\metrics\\_regression.py:918: UndefinedMetricWarning: R^2 score is not well-defined with less than two samples.\n",
      "  warnings.warn(msg, UndefinedMetricWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the test scores are non-finite: [-1.21484433 -1.45885476 -2.15621571 -1.92461061 -3.36634731         nan\n",
      "         nan         nan         nan         nan         nan         nan\n",
      "         nan         nan         nan         nan         nan         nan\n",
      "         nan]\n",
      "  warnings.warn(\n",
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the train scores are non-finite: [1.         0.63949044 0.36216616 0.23249916 0.07280011        nan\n",
      "        nan        nan        nan        nan        nan        nan\n",
      "        nan        nan        nan        nan        nan        nan\n",
      "        nan]\n",
      "  warnings.warn(\n",
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\metrics\\_regression.py:918: UndefinedMetricWarning: R^2 score is not well-defined with less than two samples.\n",
      "  warnings.warn(msg, UndefinedMetricWarning)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 1.610e+02, tolerance: 3.620e-02 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_pca_cluster_0 = helpers.train_cluster(0, pc_train_scaled, pc_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_pca_cluster_1 = helpers.train_cluster(1, pc_train_scaled, pc_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_pca_cluster_2 = helpers.train_cluster(2, pc_train_scaled, pc_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "219c4b04",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.149803</td>\n",
       "      <td>0.856950</td>\n",
       "      <td>0.104528</td>\n",
       "      <td>0.527668</td>\n",
       "      <td>2.698484</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.095890</td>\n",
       "      <td>0.445216</td>\n",
       "      <td>0.209084</td>\n",
       "      <td>1.972313</td>\n",
       "      <td>2.649144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.044122</td>\n",
       "      <td>0.852893</td>\n",
       "      <td>0.143816</td>\n",
       "      <td>1.170720</td>\n",
       "      <td>2.736672</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.059102</td>\n",
       "      <td>0.857975</td>\n",
       "      <td>0.137500</td>\n",
       "      <td>1.034857</td>\n",
       "      <td>2.747497</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.086758</td>\n",
       "      <td>0.186527</td>\n",
       "      <td>0.089450</td>\n",
       "      <td>2.273297</td>\n",
       "      <td>2.693750</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.149803  0.856950  0.104528   0.527668  2.698484\n",
       "ElasticNet      0.095890  0.445216  0.209084   1.972313  2.649144\n",
       "AdaBoost        0.044122  0.852893  0.143816   1.170720  2.736672\n",
       "RandomForest    0.059102  0.857975  0.137500   1.034857  2.747497\n",
       "KNN            -0.086758  0.186527  0.089450   2.273297  2.693750"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pca_global"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "2c8df62f",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.082149</td>\n",
       "      <td>0.911698</td>\n",
       "      <td>0.118997</td>\n",
       "      <td>0.500870</td>\n",
       "      <td>2.431879</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>-0.034722</td>\n",
       "      <td>0.429056</td>\n",
       "      <td>0.002493</td>\n",
       "      <td>1.961177</td>\n",
       "      <td>2.548807</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.156578</td>\n",
       "      <td>0.949463</td>\n",
       "      <td>-0.134714</td>\n",
       "      <td>0.648378</td>\n",
       "      <td>2.897364</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.143367</td>\n",
       "      <td>0.444697</td>\n",
       "      <td>-0.037696</td>\n",
       "      <td>2.071082</td>\n",
       "      <td>2.654796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.070271</td>\n",
       "      <td>0.027821</td>\n",
       "      <td>-0.024707</td>\n",
       "      <td>2.506757</td>\n",
       "      <td>2.640625</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.082149  0.911698  0.118997   0.500870  2.431879\n",
       "ElasticNet     -0.034722  0.429056  0.002493   1.961177  2.548807\n",
       "AdaBoost       -0.156578  0.949463 -0.134714   0.648378  2.897364\n",
       "RandomForest   -0.143367  0.444697 -0.037696   2.071082  2.654796\n",
       "KNN            -0.070271  0.027821 -0.024707   2.506757  2.640625"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pca_cluster_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "ce99784c",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>-1.355256</td>\n",
       "      <td>0.680850</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.334270</td>\n",
       "      <td>0.295962</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>-0.601424</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.002162</td>\n",
       "      <td>2.200500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.959224</td>\n",
       "      <td>0.993750</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.125000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.684504</td>\n",
       "      <td>0.806104</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1.554167</td>\n",
       "      <td>0.523333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-1.214844</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2  Test R2  Train MAE  Test MAE\n",
       "SVR            -1.355256  0.680850      NaN   1.334270  0.295962\n",
       "ElasticNet     -0.601424  1.000000      NaN   0.002162  2.200500\n",
       "AdaBoost       -0.959224  0.993750      NaN   0.125000  0.000000\n",
       "RandomForest   -0.684504  0.806104      NaN   1.554167  0.523333\n",
       "KNN            -1.214844  1.000000      NaN   0.000000  2.000000"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pca_cluster_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "6cd5bb37",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.000237</td>\n",
       "      <td>0.648274</td>\n",
       "      <td>-0.110164</td>\n",
       "      <td>0.667479</td>\n",
       "      <td>3.113053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>-0.027074</td>\n",
       "      <td>0.195380</td>\n",
       "      <td>-0.000936</td>\n",
       "      <td>1.861370</td>\n",
       "      <td>2.976239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.032755</td>\n",
       "      <td>0.928994</td>\n",
       "      <td>0.000253</td>\n",
       "      <td>0.587179</td>\n",
       "      <td>3.030796</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.029812</td>\n",
       "      <td>0.857265</td>\n",
       "      <td>0.086856</td>\n",
       "      <td>0.787269</td>\n",
       "      <td>2.899034</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.024934</td>\n",
       "      <td>0.192016</td>\n",
       "      <td>-0.006753</td>\n",
       "      <td>1.766026</td>\n",
       "      <td>3.010033</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.000237  0.648274 -0.110164   0.667479  3.113053\n",
       "ElasticNet     -0.027074  0.195380 -0.000936   1.861370  2.976239\n",
       "AdaBoost       -0.032755  0.928994  0.000253   0.587179  3.030796\n",
       "RandomForest   -0.029812  0.857265  0.086856   0.787269  2.899034\n",
       "KNN            -0.024934  0.192016 -0.006753   1.766026  3.010033"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_pca_cluster_2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "dbd9665d",
   "metadata": {
    "id": "69fIn12qwZv7"
   },
   "source": [
    "# Raw"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2de88161",
   "metadata": {},
   "source": [
    "### Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "cd026359",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:870: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "raw_train_scaled, raw_test_scaled = helpers.KM_clustering(x_train, x_test, n_clusters=3) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "233aa3ab",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_raw_global = helpers.train_all(raw_train_scaled, raw_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "f43a36aa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 6.741e-02, tolerance: 5.120e-02 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n",
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n",
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the test scores are non-finite: [-1.54302881 -0.57990449 -0.36663824 -0.24404079 -0.07183804 -0.10314303\n",
      " -0.41618374 -1.06105707         nan         nan         nan         nan\n",
      "         nan         nan         nan         nan         nan         nan\n",
      "         nan]\n",
      "  warnings.warn(\n",
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the train scores are non-finite: [1.         0.67672662 0.5854747  0.49909963 0.3866256  0.26178635\n",
      " 0.16281871 0.                nan        nan        nan        nan\n",
      "        nan        nan        nan        nan        nan        nan\n",
      "        nan]\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "results_raw_cluster_0 = helpers.train_cluster(0, raw_train_scaled, raw_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_raw_cluster_1 = helpers.train_cluster(1, raw_train_scaled, raw_test_scaled, y_train, y_test, models, plot=False)\n",
    "results_raw_cluster_2 = helpers.train_cluster(2, raw_train_scaled, raw_test_scaled, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "7a4b4831",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.253368</td>\n",
       "      <td>0.947257</td>\n",
       "      <td>-0.081751</td>\n",
       "      <td>0.296401</td>\n",
       "      <td>3.112619</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.242395</td>\n",
       "      <td>0.987845</td>\n",
       "      <td>-0.182742</td>\n",
       "      <td>0.294455</td>\n",
       "      <td>3.276554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.037996</td>\n",
       "      <td>0.882102</td>\n",
       "      <td>0.045566</td>\n",
       "      <td>1.026087</td>\n",
       "      <td>2.901213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.055412</td>\n",
       "      <td>0.886865</td>\n",
       "      <td>0.065693</td>\n",
       "      <td>0.924516</td>\n",
       "      <td>2.920500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.078388</td>\n",
       "      <td>0.372255</td>\n",
       "      <td>0.033620</td>\n",
       "      <td>2.019713</td>\n",
       "      <td>2.870833</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.253368  0.947257 -0.081751   0.296401  3.112619\n",
       "ElasticNet      0.242395  0.987845 -0.182742   0.294455  3.276554\n",
       "AdaBoost       -0.037996  0.882102  0.045566   1.026087  2.901213\n",
       "RandomForest   -0.055412  0.886865  0.065693   0.924516  2.920500\n",
       "KNN            -0.078388  0.372255  0.033620   2.019713  2.870833"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_raw_global"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "d6fa4a9d",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.159833</td>\n",
       "      <td>0.966974</td>\n",
       "      <td>0.151305</td>\n",
       "      <td>0.230943</td>\n",
       "      <td>2.701762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.142601</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.130703</td>\n",
       "      <td>0.000787</td>\n",
       "      <td>2.690017</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.077898</td>\n",
       "      <td>0.982246</td>\n",
       "      <td>0.127708</td>\n",
       "      <td>0.346582</td>\n",
       "      <td>2.735088</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.078238</td>\n",
       "      <td>0.875094</td>\n",
       "      <td>0.049753</td>\n",
       "      <td>0.996748</td>\n",
       "      <td>2.787500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.119565</td>\n",
       "      <td>0.204252</td>\n",
       "      <td>0.113675</td>\n",
       "      <td>2.395872</td>\n",
       "      <td>2.740385</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.159833  0.966974  0.151305   0.230943  2.701762\n",
       "ElasticNet      0.142601  1.000000  0.130703   0.000787  2.690017\n",
       "AdaBoost       -0.077898  0.982246  0.127708   0.346582  2.735088\n",
       "RandomForest   -0.078238  0.875094  0.049753   0.996748  2.787500\n",
       "KNN            -0.119565  0.204252  0.113675   2.395872  2.740385"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_raw_cluster_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "8c7cdcbd",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.129430</td>\n",
       "      <td>0.951410</td>\n",
       "      <td>0.149848</td>\n",
       "      <td>0.252068</td>\n",
       "      <td>2.678297</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.117679</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.012095</td>\n",
       "      <td>0.000135</td>\n",
       "      <td>2.742630</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.142137</td>\n",
       "      <td>0.982532</td>\n",
       "      <td>0.130630</td>\n",
       "      <td>0.301425</td>\n",
       "      <td>2.360295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.066718</td>\n",
       "      <td>0.841989</td>\n",
       "      <td>0.122245</td>\n",
       "      <td>0.988056</td>\n",
       "      <td>2.645679</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.018920</td>\n",
       "      <td>0.274815</td>\n",
       "      <td>-0.020384</td>\n",
       "      <td>2.020833</td>\n",
       "      <td>2.842593</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.129430  0.951410  0.149848   0.252068  2.678297\n",
       "ElasticNet      0.117679  1.000000 -0.012095   0.000135  2.742630\n",
       "AdaBoost       -0.142137  0.982532  0.130630   0.301425  2.360295\n",
       "RandomForest   -0.066718  0.841989  0.122245   0.988056  2.645679\n",
       "KNN             0.018920  0.274815 -0.020384   2.020833  2.842593"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_raw_cluster_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "544e9bfe",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.161653</td>\n",
       "      <td>0.958096</td>\n",
       "      <td>-0.275062</td>\n",
       "      <td>0.311461</td>\n",
       "      <td>3.089421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.192727</td>\n",
       "      <td>0.999970</td>\n",
       "      <td>0.048306</td>\n",
       "      <td>0.016913</td>\n",
       "      <td>2.769921</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.445217</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.347409</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>3.166667</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.267052</td>\n",
       "      <td>0.864130</td>\n",
       "      <td>-0.025740</td>\n",
       "      <td>0.989815</td>\n",
       "      <td>2.955556</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>-0.071838</td>\n",
       "      <td>0.514896</td>\n",
       "      <td>-0.785950</td>\n",
       "      <td>1.900000</td>\n",
       "      <td>3.766667</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.161653  0.958096 -0.275062   0.311461  3.089421\n",
       "ElasticNet      0.192727  0.999970  0.048306   0.016913  2.769921\n",
       "AdaBoost       -0.445217  1.000000 -0.347409   0.000000  3.166667\n",
       "RandomForest   -0.267052  0.864130 -0.025740   0.989815  2.955556\n",
       "KNN            -0.071838  0.514896 -0.785950   1.900000  3.766667"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_raw_cluster_2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "714606da",
   "metadata": {},
   "source": [
    "# Lasso Feature Selection"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "f4099e6b",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:1561: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
      "  y = column_or_1d(y, warn=True)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Lasso picked 40 variables and eliminated the other 255 variables\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 1.581e-01, tolerance: 1.130e-01\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    }
   ],
   "source": [
    "# data1 : thr=0.2, data2 : thr=0\n",
    "lasso_train, lasso_test = helpers.filter_lasso(x_train, x_test, y_train, threshold=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e1622fc",
   "metadata": {},
   "source": [
    "### Clustering"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "41b4a490",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\cluster\\_kmeans.py:870: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "lasso_train, lasso_test = helpers.KM_clustering(lasso_train, lasso_test, n_clusters=3)\n",
    "#lasso_train, lasso_test = helpers.GMM_clustering(lasso_train, lasso_test, n_clusters=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "3f06c694",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 3.719e+01, tolerance: 1.130e-01 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_lasso_global = helpers.train_all(lasso_train, lasso_test, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "5ee7e7f5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 2.644e+01, tolerance: 3.660e-02 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n",
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n",
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the test scores are non-finite: [ 0.06302853  0.13882753  0.20307204  0.20525708  0.23307547  0.18531263\n",
      "  0.15660027  0.13966241  0.12897597  0.11859472  0.08761928  0.05283715\n",
      "  0.03365372  0.00296591 -0.0297554  -0.04596888 -0.07359165 -0.10611932\n",
      "         nan]\n",
      "  warnings.warn(\n",
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\model_selection\\_search.py:953: UserWarning: One or more of the train scores are non-finite: [1.         0.77090219 0.61678195 0.51402636 0.41810622 0.36699085\n",
      " 0.33926939 0.29963057 0.26731378 0.23946598 0.19967585 0.16628561\n",
      " 0.13654427 0.11167479 0.08456986 0.05868162 0.03865291 0.02110969\n",
      "        nan]\n",
      "  warnings.warn(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 210 candidates, totalling 3150 fits\n",
      "Fitting 15 folds for each of 525 candidates, totalling 7875 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\anton\\appdata\\local\\programs\\python\\python39\\lib\\site-packages\\sklearn\\linear_model\\_coordinate_descent.py:634: ConvergenceWarning: Objective did not converge. You might want to increase the number of iterations, check the scale of the features or consider increasing regularisation. Duality gap: 5.668e+01, tolerance: 2.857e-02 Linear regression models with null weight for the l1 regularization term are more efficiently fitted using one of the solvers implemented in sklearn.linear_model.Ridge/RidgeCV instead.\n",
      "  model = cd_fast.enet_coordinate_descent(\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 15 folds for each of 30 candidates, totalling 450 fits\n",
      "Fitting 15 folds for each of 95 candidates, totalling 1425 fits\n",
      "Fitting 15 folds for each of 19 candidates, totalling 285 fits\n"
     ]
    }
   ],
   "source": [
    "results_lasso_cluster_0 = helpers.train_cluster(0, lasso_train, lasso_test, y_train, y_test, models, plot=False)\n",
    "results_lasso_cluster_1 = helpers.train_cluster(1, lasso_train, lasso_test, y_train, y_test, models, plot=False)\n",
    "results_lasso_cluster_2 = helpers.train_cluster(2, lasso_train, lasso_test, y_train, y_test, models, plot=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "509553de",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.830468</td>\n",
       "      <td>0.966649</td>\n",
       "      <td>-0.021281</td>\n",
       "      <td>0.378866</td>\n",
       "      <td>3.014441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.839701</td>\n",
       "      <td>0.967177</td>\n",
       "      <td>-0.065436</td>\n",
       "      <td>0.487145</td>\n",
       "      <td>3.038228</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>0.116188</td>\n",
       "      <td>0.899470</td>\n",
       "      <td>0.021552</td>\n",
       "      <td>0.965120</td>\n",
       "      <td>2.974360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.062492</td>\n",
       "      <td>0.894595</td>\n",
       "      <td>0.020255</td>\n",
       "      <td>0.899247</td>\n",
       "      <td>2.909421</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.131767</td>\n",
       "      <td>0.587217</td>\n",
       "      <td>0.054588</td>\n",
       "      <td>1.720430</td>\n",
       "      <td>2.840000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.830468  0.966649 -0.021281   0.378866  3.014441\n",
       "ElasticNet      0.839701  0.967177 -0.065436   0.487145  3.038228\n",
       "AdaBoost        0.116188  0.899470  0.021552   0.965120  2.974360\n",
       "RandomForest    0.062492  0.894595  0.020255   0.899247  2.909421\n",
       "KNN             0.131767  0.587217  0.054588   1.720430  2.840000"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_lasso_global"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "e258c6b1",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.377365</td>\n",
       "      <td>0.999071</td>\n",
       "      <td>0.165250</td>\n",
       "      <td>0.100012</td>\n",
       "      <td>2.352279</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.381467</td>\n",
       "      <td>0.955153</td>\n",
       "      <td>0.169096</td>\n",
       "      <td>0.536172</td>\n",
       "      <td>2.363726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.154648</td>\n",
       "      <td>0.974646</td>\n",
       "      <td>0.058729</td>\n",
       "      <td>0.402165</td>\n",
       "      <td>2.666754</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.053074</td>\n",
       "      <td>0.478338</td>\n",
       "      <td>-0.000112</td>\n",
       "      <td>1.900892</td>\n",
       "      <td>2.758333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.090138</td>\n",
       "      <td>0.484850</td>\n",
       "      <td>-0.261179</td>\n",
       "      <td>1.835294</td>\n",
       "      <td>3.042857</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.377365  0.999071  0.165250   0.100012  2.352279\n",
       "ElasticNet      0.381467  0.955153  0.169096   0.536172  2.363726\n",
       "AdaBoost       -0.154648  0.974646  0.058729   0.402165  2.666754\n",
       "RandomForest   -0.053074  0.478338 -0.000112   1.900892  2.758333\n",
       "KNN             0.090138  0.484850 -0.261179   1.835294  3.042857"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_lasso_cluster_0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "ae371803",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.530314</td>\n",
       "      <td>0.999238</td>\n",
       "      <td>0.222279</td>\n",
       "      <td>0.105163</td>\n",
       "      <td>2.615324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.534761</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.190617</td>\n",
       "      <td>0.001326</td>\n",
       "      <td>2.609755</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.011258</td>\n",
       "      <td>0.979830</td>\n",
       "      <td>0.067586</td>\n",
       "      <td>0.362090</td>\n",
       "      <td>3.164023</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>-0.002333</td>\n",
       "      <td>0.872360</td>\n",
       "      <td>-0.004063</td>\n",
       "      <td>1.122143</td>\n",
       "      <td>3.369231</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.233075</td>\n",
       "      <td>0.556591</td>\n",
       "      <td>0.052553</td>\n",
       "      <td>1.885714</td>\n",
       "      <td>3.169231</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.530314  0.999238  0.222279   0.105163  2.615324\n",
       "ElasticNet      0.534761  1.000000  0.190617   0.001326  2.609755\n",
       "AdaBoost       -0.011258  0.979830  0.067586   0.362090  3.164023\n",
       "RandomForest   -0.002333  0.872360 -0.004063   1.122143  3.369231\n",
       "KNN             0.233075  0.556591  0.052553   1.885714  3.169231"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_lasso_cluster_1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "25af85ca",
   "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>Best score</th>\n",
       "      <th>Train R2</th>\n",
       "      <th>Test R2</th>\n",
       "      <th>Train MAE</th>\n",
       "      <th>Test MAE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>SVR</th>\n",
       "      <td>0.221787</td>\n",
       "      <td>0.867763</td>\n",
       "      <td>-0.216395</td>\n",
       "      <td>0.387094</td>\n",
       "      <td>3.383254</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ElasticNet</th>\n",
       "      <td>0.154567</td>\n",
       "      <td>0.793792</td>\n",
       "      <td>0.050710</td>\n",
       "      <td>1.173765</td>\n",
       "      <td>2.899251</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>AdaBoost</th>\n",
       "      <td>-0.038448</td>\n",
       "      <td>0.981273</td>\n",
       "      <td>0.105081</td>\n",
       "      <td>0.271638</td>\n",
       "      <td>2.720330</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>RandomForest</th>\n",
       "      <td>0.063717</td>\n",
       "      <td>0.886082</td>\n",
       "      <td>-0.148547</td>\n",
       "      <td>0.823502</td>\n",
       "      <td>3.027473</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>KNN</th>\n",
       "      <td>0.063199</td>\n",
       "      <td>0.638676</td>\n",
       "      <td>-0.423183</td>\n",
       "      <td>1.365591</td>\n",
       "      <td>3.333333</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Best score  Train R2   Test R2  Train MAE  Test MAE\n",
       "SVR             0.221787  0.867763 -0.216395   0.387094  3.383254\n",
       "ElasticNet      0.154567  0.793792  0.050710   1.173765  2.899251\n",
       "AdaBoost       -0.038448  0.981273  0.105081   0.271638  2.720330\n",
       "RandomForest    0.063717  0.886082 -0.148547   0.823502  3.027473\n",
       "KNN             0.063199  0.638676 -0.423183   1.365591  3.333333"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_lasso_cluster_2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "203333d6",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}