Random Forest.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn import metrics\n",
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.model_selection import GridSearchCV"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('prediction.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Unnamed: 0</th>\n",
       "      <th>country</th>\n",
       "      <th>population</th>\n",
       "      <th>area</th>\n",
       "      <th>density</th>\n",
       "      <th>coastline_area_ratio</th>\n",
       "      <th>net_migration</th>\n",
       "      <th>infant_mortality</th>\n",
       "      <th>gdp_per_capita</th>\n",
       "      <th>literacy</th>\n",
       "      <th>...</th>\n",
       "      <th>region_BALTICS</th>\n",
       "      <th>region_C.W. OF IND. STATES</th>\n",
       "      <th>region_EASTERN EUROPE</th>\n",
       "      <th>region_LATIN AMER. &amp; CARIB</th>\n",
       "      <th>region_NEAR EAST</th>\n",
       "      <th>region_NORTHERN AFRICA</th>\n",
       "      <th>region_NORTHERN AMERICA</th>\n",
       "      <th>region_OCEANIA</th>\n",
       "      <th>region_SUB-SAHARAN AFRICA</th>\n",
       "      <th>region_WESTERN EUROPE</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>Afghanistan</td>\n",
       "      <td>31056997</td>\n",
       "      <td>647500</td>\n",
       "      <td>48.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>23.06</td>\n",
       "      <td>163.07</td>\n",
       "      <td>700.0</td>\n",
       "      <td>36.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>Albania</td>\n",
       "      <td>3581655</td>\n",
       "      <td>28748</td>\n",
       "      <td>124.6</td>\n",
       "      <td>1.26</td>\n",
       "      <td>-4.93</td>\n",
       "      <td>21.52</td>\n",
       "      <td>4500.0</td>\n",
       "      <td>86.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>Algeria</td>\n",
       "      <td>32930091</td>\n",
       "      <td>2381740</td>\n",
       "      <td>13.8</td>\n",
       "      <td>0.04</td>\n",
       "      <td>-0.39</td>\n",
       "      <td>31.00</td>\n",
       "      <td>6000.0</td>\n",
       "      <td>70.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>American Samoa</td>\n",
       "      <td>57794</td>\n",
       "      <td>199</td>\n",
       "      <td>290.4</td>\n",
       "      <td>58.29</td>\n",
       "      <td>-20.71</td>\n",
       "      <td>9.27</td>\n",
       "      <td>8000.0</td>\n",
       "      <td>97.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>Andorra</td>\n",
       "      <td>71201</td>\n",
       "      <td>468</td>\n",
       "      <td>152.1</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6.60</td>\n",
       "      <td>4.05</td>\n",
       "      <td>19000.0</td>\n",
       "      <td>100.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   Unnamed: 0          country  population     area  density  \\\n",
       "0           0     Afghanistan     31056997   647500     48.0   \n",
       "1           1         Albania      3581655    28748    124.6   \n",
       "2           2         Algeria     32930091  2381740     13.8   \n",
       "3           3  American Samoa        57794      199    290.4   \n",
       "4           4         Andorra        71201      468    152.1   \n",
       "\n",
       "   coastline_area_ratio  net_migration  infant_mortality  gdp_per_capita  \\\n",
       "0                  0.00          23.06            163.07           700.0   \n",
       "1                  1.26          -4.93             21.52          4500.0   \n",
       "2                  0.04          -0.39             31.00          6000.0   \n",
       "3                 58.29         -20.71              9.27          8000.0   \n",
       "4                  0.00           6.60              4.05         19000.0   \n",
       "\n",
       "   literacy  ...  region_BALTICS                              \\\n",
       "0      36.0  ...                                           0   \n",
       "1      86.5  ...                                           0   \n",
       "2      70.0  ...                                           0   \n",
       "3      97.0  ...                                           0   \n",
       "4     100.0  ...                                           0   \n",
       "\n",
       "   region_C.W. OF IND. STATES   region_EASTERN EUROPE                       \\\n",
       "0                            0                                           0   \n",
       "1                            0                                           1   \n",
       "2                            0                                           0   \n",
       "3                            0                                           0   \n",
       "4                            0                                           0   \n",
       "\n",
       "   region_LATIN AMER. & CARIB      region_NEAR EAST                            \\\n",
       "0                               0                                           0   \n",
       "1                               0                                           0   \n",
       "2                               0                                           0   \n",
       "3                               0                                           0   \n",
       "4                               0                                           0   \n",
       "\n",
       "   region_NORTHERN AFRICA                      \\\n",
       "0                                           0   \n",
       "1                                           0   \n",
       "2                                           1   \n",
       "3                                           0   \n",
       "4                                           0   \n",
       "\n",
       "   region_NORTHERN AMERICA                     \\\n",
       "0                                           0   \n",
       "1                                           0   \n",
       "2                                           0   \n",
       "3                                           0   \n",
       "4                                           0   \n",
       "\n",
       "   region_OCEANIA                              \\\n",
       "0                                           0   \n",
       "1                                           0   \n",
       "2                                           0   \n",
       "3                                           1   \n",
       "4                                           0   \n",
       "\n",
       "   region_SUB-SAHARAN AFRICA                   \\\n",
       "0                                           0   \n",
       "1                                           0   \n",
       "2                                           0   \n",
       "3                                           0   \n",
       "4                                           0   \n",
       "\n",
       "   region_WESTERN EUROPE                       \n",
       "0                                           0  \n",
       "1                                           0  \n",
       "2                                           0  \n",
       "3                                           0  \n",
       "4                                           1  \n",
       "\n",
       "[5 rows x 31 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Data Split 1: all of our final dataset, no scaling\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "y = df['gdp_per_capita']\n",
    "X = df.drop(['gdp_per_capita','country'], axis=1)\n",
    "\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=101)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Data Split 2: all of our final dataset, with scaling\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "sc_X = StandardScaler()\n",
    "\n",
    "X2_train = sc_X.fit_transform(X_train)\n",
    "X2_test = sc_X.fit_transform(X_test)\n",
    "y2_train = y_train\n",
    "y2_test = y_test\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Data Split 3: feature selected dataset, no scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "y3 = y\n",
    "X3 = df.drop(['gdp_per_capita','country','population', 'area', 'coastline_area_ratio', 'arable',\n",
    "                      'crops', 'other', 'climate', 'deathrate', 'industry'], axis=1)\n",
    "\n",
    "X3_train, X3_test, y3_train, y3_test = train_test_split(X3, y3, test_size=0.2, random_state=101)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Data Split 4: feature selected dataset, with scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "sc_X4 = StandardScaler()\n",
    "\n",
    "X4_train = sc_X4.fit_transform(X3_train)\n",
    "X4_test = sc_X4.fit_transform(X3_test)\n",
    "y4_train = y3_train\n",
    "y4_test = y3_test"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Random Forest"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's first try random forest with our data splits (with and without feature selection). Scaling is not going to be tested for Random Forest, since it should not affect this algorithm's performance. later we will try to improve its performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=True, ccp_alpha=0.0, criterion='mse',\n",
       "                      max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                      max_samples=None, min_impurity_decrease=0.0,\n",
       "                      min_impurity_split=None, min_samples_leaf=1,\n",
       "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                      n_estimators=200, n_jobs=None, oob_score=False,\n",
       "                      random_state=101, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf1 = RandomForestRegressor(random_state=101, n_estimators=200)\n",
    "rf3 = RandomForestRegressor(random_state=101, n_estimators=200)\n",
    "\n",
    "rf1.fit(X_train, y_train)\n",
    "rf3.fit(X3_train, y3_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf1_pred = rf1.predict(X_test)\n",
    "rf3_pred = rf3.predict(X3_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Random Forest Performance:\n",
      "\n",
      "all features, No scaling:\n",
      "MAE: 2207.804347826087\n",
      "RMSE: 3207.952113244455\n",
      "R2_Score:  0.8754543660363124\n",
      "\n",
      "selected features, No scaling:\n",
      "MAE: 2403.358695652174\n",
      "RMSE: 3549.0739491787335\n",
      "R2_Score:  0.847558627188669\n"
     ]
    }
   ],
   "source": [
    "print('Random Forest Performance:')\n",
    "\n",
    "print('\\nall features, No scaling:')\n",
    "print('MAE:', metrics.mean_absolute_error(y_test, rf1_pred))\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(y_test, rf1_pred)))\n",
    "print('R2_Score: ', metrics.r2_score(y_test, rf1_pred))\n",
    "\n",
    "print('\\nselected features, No scaling:')\n",
    "print('MAE:', metrics.mean_absolute_error(y3_test, rf3_pred))\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(y3_test, rf3_pred)))\n",
    "print('R2_Score: ', metrics.r2_score(y3_test, rf3_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "fig = plt.figure(figsize=(12, 6))\n",
    "plt.scatter(y_test,rf1_pred,color='coral', linewidths=2, edgecolors='k')\n",
    "plt.xlabel('True GDP per Capita') \n",
    "plt.ylabel('Predictions') \n",
    "plt.title('Random Forest prediction Performance (No feature selection)') \n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Optimization"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We will use grid search in order to obtain good parameters for our RF regressor. Of course our optimization here will be limited due to time and computing power constraints. The parameters we will optimize are: \n",
    "\n",
    "1. n-estimators \n",
    "2. min_samples_leaf\n",
    "3. max_features\n",
    "4. bootstrap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_param_grid = {'max_features': ['sqrt', 'auto'],\n",
    "              'min_samples_leaf': [1, 3, 5],\n",
    "              'n_estimators': [100, 500, 1000],\n",
    "             'bootstrap': [False, True]} "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_grid = GridSearchCV(estimator= RandomForestRegressor(), param_grid = rf_param_grid,  n_jobs=-1, verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=None, error_score=nan,\n",
       "             estimator=RandomForestRegressor(bootstrap=True, ccp_alpha=0.0,\n",
       "                                             criterion='mse', max_depth=None,\n",
       "                                             max_features='auto',\n",
       "                                             max_leaf_nodes=None,\n",
       "                                             max_samples=None,\n",
       "                                             min_impurity_decrease=0.0,\n",
       "                                             min_impurity_split=None,\n",
       "                                             min_samples_leaf=1,\n",
       "                                             min_samples_split=2,\n",
       "                                             min_weight_fraction_leaf=0.0,\n",
       "                                             n_estimators=100, n_jobs=None,\n",
       "                                             oob_score=False, random_state=None,\n",
       "                                             verbose=0, warm_start=False),\n",
       "             iid='deprecated', n_jobs=-1,\n",
       "             param_grid={'bootstrap': [False, True],\n",
       "                         'max_features': ['sqrt', 'auto'],\n",
       "                         'min_samples_leaf': [1, 3, 5],\n",
       "                         'n_estimators': [100, 500, 1000]},\n",
       "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
       "             scoring=None, verbose=0)"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_grid.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'bootstrap': False,\n",
       " 'max_features': 'sqrt',\n",
       " 'min_samples_leaf': 1,\n",
       " 'n_estimators': 100}"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_grid.best_params_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(bootstrap=False, ccp_alpha=0.0, criterion='mse',\n",
       "                      max_depth=None, max_features='sqrt', max_leaf_nodes=None,\n",
       "                      max_samples=None, min_impurity_decrease=0.0,\n",
       "                      min_impurity_split=None, min_samples_leaf=1,\n",
       "                      min_samples_split=2, min_weight_fraction_leaf=0.0,\n",
       "                      n_estimators=100, n_jobs=None, oob_score=False,\n",
       "                      random_state=None, verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rf_grid.best_estimator_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "rf_grid_predictions = rf_grid.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAE: 2201.6739130434785\n",
      "RMSE: 2961.5842668351033\n",
      "R2_Score:  0.8938497692158326\n"
     ]
    }
   ],
   "source": [
    "print('MAE:', metrics.mean_absolute_error(y_test, rf_grid_predictions))\n",
    "print('RMSE:', np.sqrt(metrics.mean_squared_error(y_test, rf_grid_predictions)))\n",
    "print('R2_Score: ', metrics.r2_score(y_test, rf_grid_predictions))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(12, 6))\n",
    "plt.scatter(y_test,rf_grid_predictions, linewidths=2, edgecolors='b', color='coral')\n",
    "plt.xlabel('True GDP per Capita') \n",
    "plt.ylabel('Predictions') \n",
    "plt.title('Optimized Random Forest prediction Performance (No feature selection)') \n",
    "plt.grid()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can see that the optimization process on RF regressor has not changed the performance in a noticeable manner, yet the slight change was actually to the worst, that is probably because our initial parameters were already very close to the optimum ones."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}