competition-procedure.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Begining exploration of the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from cleaning.clean_func import *\n",
    "from pipe.fitting_models import *\n",
    "\n",
    "from sklearn import preprocessing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "\n",
    "### Features\n",
    "- id: only for test & sample submission files, id for prediction sample identification\n",
    "- price: price in USD\n",
    "- carat: weight of the diamond\n",
    "- cut: quality of the cut (Fair, Good, Very Good, Premium, Ideal)\n",
    "- color: diamond colour, from J (worst) to D (best)\n",
    "- clarity: a measurement of how clear the diamond is (I1 (worst), SI2, SI1, VS2, VS1, VVS2, VVS1, IF (best))\n",
    "- x: length in mm\n",
    "- y: width in mm\n",
    "- z: depth in mm\n",
    "- depth: total depth percentage = z / mean(x, y) = 2 * z / (x + y) (43--79)\n",
    "- table: width of top of diamond relative to widest point (43--95)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>carat</th>\n",
       "      <th>cut</th>\n",
       "      <th>color</th>\n",
       "      <th>clarity</th>\n",
       "      <th>depth</th>\n",
       "      <th>table</th>\n",
       "      <th>x</th>\n",
       "      <th>y</th>\n",
       "      <th>z</th>\n",
       "      <th>price</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.50</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>D</td>\n",
       "      <td>VS2</td>\n",
       "      <td>62.3</td>\n",
       "      <td>55.0</td>\n",
       "      <td>5.11</td>\n",
       "      <td>5.07</td>\n",
       "      <td>3.17</td>\n",
       "      <td>1845</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.54</td>\n",
       "      <td>Good</td>\n",
       "      <td>I</td>\n",
       "      <td>VS1</td>\n",
       "      <td>63.6</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7.30</td>\n",
       "      <td>7.33</td>\n",
       "      <td>4.65</td>\n",
       "      <td>10164</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.32</td>\n",
       "      <td>Very Good</td>\n",
       "      <td>J</td>\n",
       "      <td>SI2</td>\n",
       "      <td>61.7</td>\n",
       "      <td>60.0</td>\n",
       "      <td>6.95</td>\n",
       "      <td>7.01</td>\n",
       "      <td>4.31</td>\n",
       "      <td>5513</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.20</td>\n",
       "      <td>Ideal</td>\n",
       "      <td>I</td>\n",
       "      <td>SI1</td>\n",
       "      <td>62.1</td>\n",
       "      <td>55.0</td>\n",
       "      <td>6.83</td>\n",
       "      <td>6.79</td>\n",
       "      <td>4.23</td>\n",
       "      <td>5174</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.73</td>\n",
       "      <td>Premium</td>\n",
       "      <td>I</td>\n",
       "      <td>SI1</td>\n",
       "      <td>61.2</td>\n",
       "      <td>60.0</td>\n",
       "      <td>7.67</td>\n",
       "      <td>7.65</td>\n",
       "      <td>4.69</td>\n",
       "      <td>10957</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    carat        cut color clarity  depth  table     x     y     z  price\n",
       "id                                                                       \n",
       "0    0.50      Ideal     D     VS2   62.3   55.0  5.11  5.07  3.17   1845\n",
       "1    1.54       Good     I     VS1   63.6   60.0  7.30  7.33  4.65  10164\n",
       "2    1.32  Very Good     J     SI2   61.7   60.0  6.95  7.01  4.31   5513\n",
       "3    1.20      Ideal     I     SI1   62.1   55.0  6.83  6.79  4.23   5174\n",
       "4    1.73    Premium     I     SI1   61.2   60.0  7.67  7.65  4.69  10957"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "diamonds = pd.read_csv(\"train.csv\",index_col=\"id\")\n",
    "diamonds.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Importing ML models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import linear_model\n",
    "from sklearn.model_selection import GridSearchCV\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.linear_model import ElasticNet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "1st MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_1(diamonds.drop(columns=[\"price\"]))\n",
    "scaler = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 1002.7527392109839\n",
      "Choosen one best 1002.7527392109839 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  9.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 652.7200024388085\n",
      "Choosen one best 652.7200024388085 1002.7527392109839 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed: 11.8min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 721.2007310622586\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "KNeighborsRegressor(weights='distance')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func,rmse_1 = best_func(X,y)\n",
    "sel_func.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "652.7200024388085\n",
      "KNeighborsRegressor(weights='distance')\n"
     ]
    }
   ],
   "source": [
    "print(rmse_1)\n",
    "print(sel_func)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 14.457395849648082\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import r2_score,mean_squared_error\n",
    "y_pred=sel_func.predict(X)\n",
    "rmse = mean_squared_error(10**y.values, 10**y_pred, squared=False)\n",
    "print(\"RMSE (test) =\",rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_1(diamonds_test)\n",
    "Xf=scaler.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_1_pred.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "2nd MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:41: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:42: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:43: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:50: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[f\"{col1}*{col2}\"] = diamonds[col1]*diamonds[col2]\n"
     ]
    }
   ],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_2(diamonds.drop(columns=[\"price\"]))\n",
    "scaler2 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler2.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#(min_samples_leaf=5, min_samples_split=5, n_estimators=50, n_jobs=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 6193.006723698375\n",
      "Choosen one best 6193.006723698375 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:   53.5s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 651.1109536693785\n",
      "Choosen one best 651.1109536693785 6193.006723698375 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  7.1min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 571.3367727564453\n",
      "Choosen one best 571.3367727564453 651.1109536693785 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func2,rmse_2 = best_func(X,y)\n",
    "sel_func2.fit(X,y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "571.3367727564453\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_2)\n",
    "print(sel_func2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:41: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:42: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:43: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:50: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[f\"{col1}*{col2}\"] = diamonds[col1]*diamonds[col2]\n"
     ]
    }
   ],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_2(diamonds_test)\n",
    "Xf=scaler2.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func2.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_2_pred.csv\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "3rd MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_3(diamonds.drop(columns=[\"price\"]))\n",
    "scaler3 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler3.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 5610.2217587438945\n",
      "Choosen one best 5610.2217587438945 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  7.8min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 692.7519036665183\n",
      "Choosen one best 692.7519036665183 5610.2217587438945 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  6.5min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 661.788541347771\n",
      "Choosen one best 661.788541347771 692.7519036665183 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_estimators=75,\n",
       "                      n_jobs=-1)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func3,rmse_3 = best_func(X,y)\n",
    "sel_func3.fit(X,y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "661.788541347771\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_estimators=75,\n",
      "                      n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_3)\n",
    "print(sel_func3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_3(diamonds_test)\n",
    "Xf=scaler3.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func3.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_3_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "4th MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:68: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:69: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:70: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:72: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"all_cat\"]=diamonds[\"color\"]*diamonds[\"cut\"]*diamonds[\"clarity\"]\n"
     ]
    }
   ],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_4(diamonds.drop(columns=[\"price\"]))\n",
    "scaler4 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler4.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 28155.65257380409\n",
      "Choosen one best 28155.65257380409 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:   35.9s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 669.5573221592437\n",
      "Choosen one best 669.5573221592437 28155.65257380409 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  3.6min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 562.4066484968095\n",
      "Choosen one best 562.4066484968095 669.5573221592437 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_estimators=75,\n",
       "                      n_jobs=-1)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func4,rmse_4 = best_func(X,y)\n",
    "sel_func4.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "562.4066484968095\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_estimators=75,\n",
      "                      n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_4)\n",
    "print(sel_func4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:68: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:69: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:70: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:72: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"all_cat\"]=diamonds[\"color\"]*diamonds[\"cut\"]*diamonds[\"clarity\"]\n"
     ]
    }
   ],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_4(diamonds_test)\n",
    "Xf=scaler4.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func4.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_4_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "5th MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_5(diamonds.drop(columns=[\"price\"]))\n",
    "scaler5 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler5.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 1551.9736968689926\n",
      "Choosen one best 1551.9736968689926 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  8.3min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 680.2852679663085\n",
      "Choosen one best 680.2852679663085 1551.9736968689926 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  7.2min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 705.3429143862282\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "KNeighborsRegressor(weights='distance')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func5,rmse_5 = best_func(X,y)\n",
    "sel_func5.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "680.2852679663085\n",
      "KNeighborsRegressor(weights='distance')\n"
     ]
    }
   ],
   "source": [
    "print(rmse_5)\n",
    "print(sel_func5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_5(diamonds_test)\n",
    "Xf=scaler5.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func5.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_5_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "6th MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:89: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:90: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:91: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:98: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[f\"{col1}*{col2}\"] = diamonds[col1]*diamonds[col2]\n"
     ]
    }
   ],
   "source": [
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_6(diamonds.drop(columns=[\"price\"]))\n",
    "scaler6 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler6.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 1937.121084567295\n",
      "Choosen one best 1937.121084567295 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:   54.4s finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 820.450643166402\n",
      "Choosen one best 820.450643166402 1937.121084567295 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  7.4min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 673.9502776468439\n",
      "Choosen one best 673.9502776468439 820.450643166402 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=25, n_jobs=-1)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func6,rmse_6 = best_func(X,y)\n",
    "sel_func6.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "673.9502776468439\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=25, n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_6)\n",
    "print(sel_func6)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:89: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"color\"]=diamonds[\"color\"].replace(color)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:90: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"cut\"]=diamonds[\"cut\"].replace(cut)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:91: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[\"clarity\"]=diamonds[\"clarity\"].replace(clarity)\n",
      "/home/ordovas/IRONHACK/mini-projects/kaggle-competition/cleaning/clean_func.py:98: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame.\n",
      "Try using .loc[row_indexer,col_indexer] = value instead\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  diamonds[f\"{col1}*{col2}\"] = diamonds[col1]*diamonds[col2]\n"
     ]
    }
   ],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_6(diamonds_test)\n",
    "Xf=scaler6.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func6.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_6_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "7th MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [],
   "source": [
    "from cleaning.clean_func import cleaning_7\n",
    "\n",
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_7(diamonds.drop(columns=[\"price\"]))\n",
    "scaler7 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler7.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 5797.372723710179\n",
      "Choosen one best 5797.372723710179 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  2.3min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 704.1593339923847\n",
      "Choosen one best 704.1593339923847 5797.372723710179 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  6.4min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 543.604606674903\n",
      "Choosen one best 543.604606674903 704.1593339923847 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func7,rmse_7 = best_func(X,y)\n",
    "sel_func7.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "543.604606674903\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_7)\n",
    "print(sel_func7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_7(diamonds_test)\n",
    "Xf=scaler7.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func7.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_7_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe8283f05b0>]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y,sel_func7.predict(X))\n",
    "plt.plot([2.5,4.25],[2.5,4.25],color=\"red\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fe822604250>]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(10**y,10**sel_func7.predict(X))\n",
    "plt.plot([0,17800],[0,17800],color=\"red\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 898.3855961904379\n",
      "RMSE (test) = 219.85899545527053\n"
     ]
    }
   ],
   "source": [
    "rmse = mean_squared_error(10**y[3.8<y], 10**sel_func7.predict(X)[3.8<y], squared=False)\n",
    "print(\"RMSE (test) =\",rmse)\n",
    "rmse = mean_squared_error(10**y[y<3.8], 10**sel_func7.predict(X)[y<3.8], squared=False)\n",
    "print(\"RMSE (test) =\",rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 1116.8115677235785\n",
      "RMSE (test) = 296.1324764743617\n"
     ]
    }
   ],
   "source": [
    "rmse = mean_squared_error(10**y[4<y], 10**sel_func7.predict(X)[4<y], squared=False)\n",
    "print(\"RMSE (test) =\",rmse)\n",
    "rmse = mean_squared_error(10**y[y<4], 10**sel_func7.predict(X)[y<4], squared=False)\n",
    "print(\"RMSE (test) =\",rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "42.44855844956533"
      ]
     },
     "execution_count": 95,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sum((y-sel_func7.predict(X))**2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "8th MODEL"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_8(diamonds.drop(columns=[\"price\"]))\n",
    "scaler8 = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X=scaler8.transform(X_raw)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 1015.581112283209\n",
      "Choosen one best 1015.581112283209 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  9.2min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 735.7571126804845\n",
      "Choosen one best 735.7571126804845 1015.581112283209 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed: 20.1min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 581.8298074127849\n",
      "Choosen one best 581.8298074127849 735.7571126804845 True\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sel_func8,rmse_8 = best_func(X,y)\n",
    "sel_func8.fit(X,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "581.8298074127849\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)\n"
     ]
    }
   ],
   "source": [
    "print(rmse_8)\n",
    "print(sel_func8)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "Xf_raw=cleaning_8(diamonds_test)\n",
    "Xf=scaler8.transform(Xf_raw.values)\n",
    "y_prediction = 10**sel_func8.predict(Xf)\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_8_pred.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Testing dividing in 2 samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n",
    "y=y_processing(diamonds[\"price\"])\n",
    "X_raw=cleaning_8(diamonds.drop(columns=[\"price\"]))\n",
    "scaler_2parts = preprocessing.StandardScaler().fit(X_raw.values)\n",
    "X_raw1=cleaning_8(diamonds[6.75<diamonds[\"x\"]].drop(columns=[\"price\"]))\n",
    "X_raw2=cleaning_8(diamonds[diamonds[\"x\"]<6.75].drop(columns=[\"price\"]))\n",
    "X1=scaler_2parts.transform(X_raw1)\n",
    "X2=scaler_2parts.transform(X_raw2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 2.442568270133245e+26\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 1457.464515614\n",
      "Choosen one best 1457.464515614 1000000 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  1.0min finished\n",
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed:  8.1min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 1130.1663877506726\n",
      "Choosen one best 1130.1663877506726 1457.464515614 True\n"
     ]
    }
   ],
   "source": [
    "sel_func_2parts_1,rmse__2parts_1 = best_func(X1,y[6.75<diamonds[\"x\"]])\n",
    "#sel_func_2parts_1.fit(X1,y[6.75<diamonds[\"x\"]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "LinearRegression\n",
      "RMSE (test) = 516.1122396877047\n",
      "Choosen one best 516.1122396877047 1000000 True\n",
      "KNN\n",
      "Fitting 5 folds for each of 18 candidates, totalling 90 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done  90 out of  90 | elapsed:  6.7min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 414.97039173371763\n",
      "Choosen one best 414.97039173371763 516.1122396877047 True\n",
      "Forest\n",
      "Fitting 5 folds for each of 336 candidates, totalling 1680 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=1)]: Using backend SequentialBackend with 1 concurrent workers.\n",
      "[Parallel(n_jobs=1)]: Done 1680 out of 1680 | elapsed: 15.8min finished\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 322.39832060242776\n",
      "Choosen one best 322.39832060242776 414.97039173371763 True\n"
     ]
    }
   ],
   "source": [
    "sel_func_2parts_2,rmse__2parts_2 = best_func(X2,y[diamonds[\"x\"]<6.75])\n",
    "#sel_func_2parts_2.fit(X2,y[diamonds[\"x\"]<6.75])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1130.1663877506726\n",
      "RandomForestRegressor(max_features='log2', min_samples_leaf=5,\n",
      "                      min_samples_split=5, n_jobs=-1)\n",
      "322.39832060242776\n",
      "RandomForestRegressor(min_samples_leaf=5, min_samples_split=5, n_jobs=-1)\n",
      "569.4520991194843\n"
     ]
    }
   ],
   "source": [
    "print(rmse__2parts_1,)\n",
    "print(sel_func_2parts_1)\n",
    "print(rmse__2parts_2)\n",
    "print(sel_func_2parts_2)\n",
    "print(np.sqrt(  (rmse__2parts_1*rmse__2parts_1*len(y[6.75<diamonds[\"x\"]])/len(y))\n",
    "      +(rmse__2parts_2*rmse__2parts_2*len(y[diamonds[\"x\"]<6.75])/len(y)) )  )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "575.9261994985208"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "331690.98726881**0.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE (test) = 467.73047327254767\n"
     ]
    }
   ],
   "source": [
    "yp=y.copy()\n",
    "yp[6.75<diamonds[\"x\"]]=sel_func_2parts_1.predict(X1)\n",
    "yp[diamonds[\"x\"]<6.75]=sel_func_2parts_2.predict(X2)\n",
    "rmse = mean_squared_error(10**y, 10**yp, squared=False)\n",
    "print(\"RMSE (test) =\",rmse)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fc7467cde50>]"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAD4CAYAAADiry33AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8vihELAAAACXBIWXMAAAsTAAALEwEAmpwYAAA4l0lEQVR4nO3deXhTVfrA8e/bNEDLVlaBQsEFUQEFRUBxAWRVRERFBNxGxQ1xRUEZAUVB+psRVBRxd0RHRqWilk1B2QRZyr4oIAIFZS1rgTY9vz+aYEiTJm3vTdP0/TwPj+3NzT0nmcyb03Pf8x4xxqCUUip6xRR3B5RSStlLA71SSkU5DfRKKRXlNNArpVSU00CvlFJRLra4O+BP9erVTYMGDYq7G0opVWIsW7ZsrzGmhr/HIjLQN2jQgKVLlxZ3N5RSqsQQkT8CPaZTN0opFeU00CulVJTTQK+UUlFOA71SSkU5DfRKKRXlQs66EREHsBRIN8Z083msL/CM+9cjwIPGmJXux7YChwEXkG2MaWFBv5VSKiqkpKWTPGMjOzMyqZMQx6DOjejRPNHSNgqSXvkosB6o5Oex34GrjTEHRKQrMBFo5fV4O2PM3sJ3UymlokdKWjrPfLmKE9k5px1Pz8jksc9XMCp1HYuf62hZeyEFehGpC1wHvAQ84fu4MWah16+LgLqW9E4ppUogzyg9PSMz4Dl1D/5F2eyTbK5WL89jfx0+Scd//8isJ9pa0p9QR/RjgaeBiiGcew8wzet3A8wUEQO8bYyZWKAeKqVUhEtJS2fEN2s5cCwr6LlicuiXlsrgHz9kda1z6N1ntN/zftt91LL+BQ30ItIN2G2MWSYibYOc247cQH+F1+E2xpidIlITmCUiG4wxc/08tz/QHyApKSn0V6CUUhbxHok7RHAZgwh49mdKiHPS7aLazNmwJ9/ReiD1D+xkzLTXaLV9DT+deTFDugyw+BX4F8qIvg3QXUSuBcoBlUTkE2NMP++TRORC4F2gqzFmn+e4MWan+7+7RWQK0BLIE+jdI/2JAC1atNBtr5RStvF3AxRgyFerycxyAeByR3fvTfgyMrP4ZNG2ArcXk+Pi7qVTeWreJ5x0xPLUtY/xRZNrQKToLyYEQQO9MWYIMATAPaJ/yk+QTwK+Am43xvzqdbw8EGOMOez+uRPwgmW9V0qpAhqasppJi7bhid/pGZkM+mIlWS57xpfn7N1Gcuo4mu/ayKxzWvFcp4fYXbGaLW0FUuiiZiLyAIAxZgLwPFANeFNyv6E8aZRnAFPcx2KBT40x04vaaaWUyk+gEXugeXQ7grwjx8X9i7/k0QWfcrRMPAOvH8TU868K2yjem0Ti5uAtWrQwWr1SKVUYKWnpp03BADhjhBzAlROeeHfe7t9JTh1L0782822jKxjW8QH2lU8o8HW2jr4u5HNFZFmgdUoRWaZYKaUKK3nGxtOCPEBWmAK805XFwz9P5uGfJ5NRriIP9BjC9EZtwtJ2fjTQK6Wiys5CZMNYocmfm0hOHcv5e7byVeN2vHDNfWTE+VtfGn4a6JVSUaVOQlyhUh8Lq2z2SR5d8Cn9F3/F3vIJ/OOm55l9TssiX9fKmXwtaqaUiiqDOjcizukIS1sXp6/nuw8G8tCiL/iiaQc63fOmJUEe4NVbm1lyHdARvVIqCvhb6GSnclnHeWruf/jH0qnsrFSdfr1eZP6ZzS27voClhc000CulSrSUtPTT8uDtDvKttq3mlWmv0SBjFx83v45Xrr6To2XjLW3D6leggV4pVSKFUjjMSuVPHOOZnz7ijrTv2JpQm1tvG8XipKa2tJWYEGfp9TTQK6VKHH+58na64vc0Rk9/nTqH9vBuixv4v6tu57iznC1txTkdpxZ4WUUDvVKqxAj3KL7S8SM8O+d9eq+ayaaqdbm53xiWJ55va5tlY63PkdFAr5SKCMF2Wgr3KL7d5iW8PP0Nah49wJutb2Zcmz6ciC1je7sZmVkM+Wo1YN0NWQ30Sqli5xvE0zMy8wQ7fyte7ZCQeYjnf3iHnmvnsL5GA/r3HMrq2g1tay/eGcOxrNN3msrMcpE8Y6MGeqVU9Bjxzdo8Qdw32IVjxWvnjQsZOetNEjIPM7bNbYy/rBdZDqetbfoGeQ8rp6c00CulilVKWnrAnZnSMzJp/sJMjLE+5dBbtaMZjJg1gW4b57PmjLO5o9cLrK95lo0tBuewsMqlBnqlVLFKnrEx38dD2Z6v0Iyh+/q5DP/+bcqfPMaYq+5gYsueZDuKPzRauR6g+F+NUioqBLuZGkg469J4q3l4HyNnvUWn3xaxvE4jBnV9jM3V827UXVyszKXXQK+UKrJQbqYGep5g77RMHsZw85of+OcP71DWlcWL7e7hgxbdyYkJT32cUAhYmkuvgV4pVWT+MmJCyRxJnrExrEG+zqHdvDx9PG1/X8biuo15putAtla1rqaMFQTo2zpJa90opSJLoIyYYJky4aodLyaH21bOYMic94kxhn92fIBPml+Lkcgr4Pvqrc0sDfKggV4pZYFANeDr5DPPnJKWbmeXTqmX8SevTHuNy7etYl79Zgzp+gg7Kp8RlrYLKjEhzvIgDxrolVIWGNS5UZ5Vq741W7xv1sb5WSRkNTE53LnsW56e+xHZ4uCZLo/w+YWdbNmcOyHOSfmysezMyCQh3smR49mnbV/ouQ+R3/0IO2rceIQc6EXEASwF0o0x3XweE2AccC1wDLjLGLPc/VgX92MO4F1jzGiL+q6UihDeq1c9WTftzqtB8oyNPP75ijzBz+4gf+b+dMakjuPS9HXMPqsFz3V+mF2VatjSVpzTwfDujfOUa/CXgeR9vHKcExHIOJZVoCylwhATYq6miDwBtAAq+Qn01wKPkBvoWwHjjDGt3F8OvwIdgR3AEuA2Y8y6/Npq0aKFWbp0aUFfi1IqQoS7Lo2HI8fFPUtSeGL+JI7HlmHENf2Z0rid5aP48mUcHDvpsj1AF4SILDPGtPD3WEgjehGpC1wHvAQ84eeUG4CPTe63xiIRSRCR2kADYJMxZov7Ov91n5tvoFdKlWzhqkvjreGeP0ieNpZmu35jRsPWDO30EHsqVLW8nX6tkxjZw5469HYJdepmLPA0UDHA44nAdq/fd7iP+Tveyt8FRKQ/0B8gKSkpxG4ppSJRuLJpAGJd2Tyw+AsGLvgvh8vGM6D703x73pWWj+LjnA5G9WwaEaP3ggoa6EWkG7DbGLNMRNoGOs3PMZPP8bwHjZkITITcqZtg/VJKRRbv+eeYMOzbCnDBX1tITh1L491bmHr+VQzvcD/74ysX+brxzhhe7nlhoVb6RqJQRvRtgO7uefhyQCUR+cQY08/rnB2A99rhusBOoEyA40qpKOI7J293kC+TncXDP3/OQ4v+x4G4SvS/8TlmnnuZJdeOEXi554X0aJ5YYgO7r6CB3hgzBBgC4B7RP+UT5AGmAgPcc/CtgIPGmF0isgdoKCJnAulAb6CPdd1XSoVDsDo24ZyTv2jnRsZMG0ejvdv4osk1vNj+Xg7GBZpVLpg4Zwyj3EE+mhQ6j15EHgAwxkwAUsnNuNlEbnrl3e7HskVkADCD3PTK940xa4vaaaVU+PirY/P45yt47PMVJLqDfjgKk5XNOsHj8ydx35IU/qpQlbtuHsaPZ19q2fVL4k3WUIWcXhlOml6pVORoM3p2voE8zumwfTR/yY51jJk2jrP3p/PpRV0Y1e5uDpctb2kbiQlxLBjc3tJrhlOR0yuVUqVXsAwaO4N83MnjPD33I+5c9i3plWvS59aRLGzQzJa2wpkpFG4a6JVS+QpUx8Zul/2xilemjSPp4F98cMn1JF91B8fKWFej3Vd+dXlKusgr3aaUiiiDOjfC6bC+PkwgFU4cY+SM8Xz232dxxcRwS5/RjOhwv61B3ur675FGR/RKlWL+smkgb80aV0547uVdvWUZL09/g1pH9vF2y568ekUfjjvLWdqGb2ExO+q/Rxq9GatUKeWvHo0jRsIW1L1VOn6Ef/7wLres+Z7fqtVj0LWPsaKO9SNsT5ZQtCyE8qY3Y5VSefjLfS+OIN/ht8W8NHM81Y5m8MZlvXjt8ts4Geu0pa2dGZlRtRAqVBrolSqlimtTbo8qxw4y7IeJ9Fj3E+trNOAfNz3P2lrn2NpmNN9wzY8GeqVKKUeY6tH4c+2G+bww6y0qHT/Kv67oy4TWN5PlsGcU72Hnxh6RTgO9UqVUcQT56kcP8MLMt7j214WsrNWQvr0fZWONBra0Fe+MoUr5slE3F18YGuiVKqXCOqI3hh7rfmTY9xOJzzrOqLZ38e6lN+KKcdjSXJzTwcsltKSwHTTQK1VKhSvIn3F4Ly/NGE+HzUtYVuc8nr72UTZXqxf8iSGIEejTKokW9atGZSaNVTTQK1UKpaSl29+IMfRaNYuhs9/FmePihfb38eEl3cixYBTvbxMQDeyBaaBXqhRJSUtnxDdrOXAsy9Z2Eg/uZtT017lqaxqL6jXh6a6Psq1KbUuuLcBNl5S+FMmi0ECvVJQIVjM+HBt2i8mhb9o0Bv/0IQBDOz3EpGZdMGJdtRUDzNmwx7LrlQYa6JWKAv5qxg/5ajXw95SG3ZuD1D+wk1emvUbr7WuY26A5Q7o8Qnrlmra0FajSZLAvu9JKA71SUcBfEM/McvHk5JVAbrC3a4FUTI6Lu5d9w1Nz/0OWI5ZBXQfyv6YdLd+c25u/hU+hfNmVVhrolYoCgUa4LmMY8tVqRqWus6Xds/dtZ0zqOC7ZuYHvz76U5zo/zF8Vq1vahm8RskALnwJ92SXP2KiBvrg7oJQquvxqxmdmuSyfsnHkuLjvlyk8Pn8Sx5zleLTbk3x9QVvLRvEJcU6Gd29Mj+aJIU/HBPqyi+YNRUKlgV6pKDCocyPbb7R6NNqzleTUsVz45yZSz72c5zs9yN7yVSy5tgi82qtZnrTJUEbkgb7sSmt9G29BA72IlAPmAmXd539hjBnmc84goK/XNc8Hahhj9ovIVuAw4AKyA5XRVEoVTdnYGFsDvdOVxYOLvmDAws85VK48D94wmGnnXWHZ9eOcDm66JJHkGRt5/PMVBb6Z6u/LrjTXt/EWyoj+BNDeGHNERJzAfBGZZoxZ5DnBGJMMJAOIyPXA48aY/V7XaGeM2Wtlx5UqbfxNYfxv6TYWbN4f/MlF1OTPTSSnjuX8PVuZckFbXrjmPg7EV7bs+p7c+C+XpRf6Zqp3dlE4s25KQqZPgTYeEZF4YD7woDFmcYBzPgXmGGPecf++FWhRkECvG48odbpw5MD7Uzb7JAMXfMb9i79kX/kEnuv0MN83bGVLW1XinX4XciUmxLFgcHtb2iwqf/+7+Fu1Gw5F3nhERBzAMuAcYHw+QT4e6AIM8DpsgJkiYoC3jTETAzy3P9AfICkpKZRuKRVV8hsZ2p0D70/z9A2MmTaOhvu2M7lpB0a2v5dD5SrY1l6g1bqRfDO1pGT6hBTojTEuoJmIJABTRKSJMWaNn1OvBxb4TNu0McbsFJGawCwR2WCMmeunjYnARMgd0Rf0hShVUvkrS+A7bRHOYFcu6zhPzJvEPUu/5s8K1bjjlhHMPeuSsLXvK5JvppaUTJ8CrUs2xmQAP5I7avenN/CZz3N2uv+7G5gCtCxoJ5WKVp4//f2NZj0jQwhfsGu5fQ3TPniE/kum8GmzLnS+Z7zlQT5QAmZCnJM45+kFzyL9Zmqg/10i7cspaKAXkRrukTwiEgd0ADb4Oa8ycDXwtdex8iJS0fMz0Anw95eAUlEjJS2dNqNnc+bg72gzena+lSKDTcl4RoaDOjfKEwStFH8yk+GzJjD508HEGMNtvV/mn50e4kjZeMvb6ts6yW9AH969MaN6NiUxIQ4hd26+OOa6C8Lf/y6R+OUUytRNbeAj9zx9DDDZGPOtiDwAYIyZ4D7vRmCmMeao13PPIHeqx9PWp8aY6Zb1XqkIU9Bl+MH+xPceGWa5cizs6d/abF3B6Omvk3hwN+9f0p3kq+4gs0w5W9qKc8YwskfTfOvHR3Jg95Vfpk8kZeMUKOsmXDTrRpVUbUbP9rtox5M5MjRlNZ8t3o7LGBwilIkVMrP8B3BP9gbAE5NXkGPx/1UrnjjKkDnv02flDDZXTeTpro+yrO4F1jbixRkjJN9yUYkK5IVVHNk4Rc66UUqFJr+bc0NTVvPJom2njrmMITMrcPTOzHLx2OcrrO4iAG03L2HU9DeoefQAE1rdxKtt+nDCWdbydjx1ahIjNL/cLpGWjaOBXikL5bcM/7PF24uhR6ernHmY52e/w01rZrOxehL393yOVbXPtaUt73o1pU2kZeNYtxuAUirfm3Nh24g7gE6//sys9x6i+7qfGHd5b66/c5xtQR5gxbBOpTLIQ+Rl42igV8pCPZonBswccdhYnz0/VY8d5PWvX2HilJfYXaEqN9zxKq9e2Y+TsU5b220QQtZRtIq0bBy9GauUDTwZF+kZmThEcBlDnDMm4I1XWxjD9evnMvz7t6lw8hivXX4bb7e6iWxHeGdsi6skQHELd9ZNfjdjNdArZbH86tLEAAiWZ9D4qnFkPyNnvknn3xaxova5DOr6KL/VqG9vo/mI5Ho10UKzbpQKo/wWQeUAcbE2juyNoefa2Tz/wzuUyz7JS23/wfuX3oArxr7FVqGItJIApY0GeqUsFmxvVruCfO1De3h5xhu027KMX+pewDNdH+X3qpExXRJpJQFKGw30SlnMMycfNsbQe+UMnpvzHg6Tw7AO9/PxxddhJHy5FoFKDENklgQobTTQK2WxcAb5uhl/Mnr661zxx0oW1L+QwV0Gsj2hVtjah9Pn31PS0hk+dS0ZmblBv0q8k2HXl85c+kiigV6pQgiUUZGSlh6WEb2YHG5f/h3P/PQROSIM6TyAzy7qbNnm3KHyHq37uwl9PJxZRiogDfRKFVCgwmVL/9jPl8vSbQ/yDfanM2baOFruWMePZ17CkC4D2FWphq1tekoZePMdrUfasn/1Nw30SvlR0N2eMrNcp9WxsUNMjot/LP2ap+Z9wgmHkyevfZwvm7QPyyje31dXfJnY0wJ4pC37V3/TQK+Uj2ClhosjcJ2zdxvJqeNovmsjs85pxXOdHmJ3xWqWt1OQaSff9yG/Oj+qeGkJBKV85DcFAeENXLGubB76eTLffTiQ+hm7eOT6QdzXc6gtQT4GuK1VvTxL9wP9veD7PkTasn/1Nx3RK+Uj0Ig9PSOTlLR0BnVuFHDlq5XO372F5NRxNPlrM9+edyXDOtzPvvIJtrVXOd7pd1OQdufV4Mtl6Xlqq/sG8Pw24ShNImnDEQ8tgaCUj0Cbh8Dpm4F4pxFayenKYsDCyTy0aDIZcRUZ2vEhZjS63PJ2fAnw++jr/D4WicErEhXHhiMeWgJBqQLIb8TumcJpd14NDtoQ5Jvu+o3k1LGct/cPvmzcjhevuY+MuEqWt+NPflNSPZonamAPQaRmHmmgV8qH5/+QgXZ3Ss/ItDzDpmz2SR6b/yn9f/mK3eWrcPfNw5hz9qWWtpEfnUu3RqRmHgW9GSsi5UTkFxFZKSJrRWSEn3PaishBEVnh/ve812NdRGSjiGwSkcFWvwCl7NCjeSKJYbrpevGO9aR+MJAHF3/B5KYd6HTvm2EL8r4181XRRNqGIx6hjOhPAO2NMUdExAnMF5FpxphFPufNM8Z08z4gIg5gPNAR2AEsEZGpxph1VnReKbukpKVz7GS2rW3EnTzOU/P+w91Lp7KzUg369XqR+Wc2t7VNb/1aJzGyR9OwtVca+Jv2i4S/loIGepN7t/aI+1en+1+od3BbApuMMVsAROS/wA2ABnoVsfKrJ2+V1ttW8cq016if8ScfXXwdY666k6Nl421rz5tDhNta1dMgb4NIzTwKaY7ePTJfBpwDjDfGLPZz2mUishLYCTxljFkLJALeOyLvAFoFaKM/0B8gKSkp5BeglNXyqydfVOVPHGPwTx9ye1oqWxNq06vPaH6p18SWtrxZvVG3ZuEEFok3rkMK9MYYF9BMRBKAKSLSxBizxuuU5UB99/TOtUAK0BD/ay38/jVgjJkITITc9MqQX4FShZBfoApWT76wrvx9OaOmv06dQ3t559Ie/OvKfhx3lrO0DadDuPXSeszZsMe2IBxs5XBJF41fYgXKujHGZIjIj0AXYI3X8UNeP6eKyJsiUp3cEXw9r0vUJXfEr1Sx8ReoBv1vJSO+WRuwpnpRVDp+hOdmv8etq2exqWpdbu43huWJ51veTmKYglKkphBaIVq/xIIGehGpAWS5g3wc0AF4xeecWsBfxhgjIi3JzebZB2QADUXkTCAd6A30sfYlKFUw/gJVVo6xJci33/QLL894g+pHMxjf+hZea3MbJ2LLWNpGuG+qRmoKoRWi9UsslBF9beAj9zx9DDDZGPOtiDwAYIyZANwMPCgi2UAm0Nt9EzdbRAYAMwAH8L577l6pYmPX1Iy3hMxDDPt+Ijeu+5H1NRpw703Ps6bWOZa2IUDfYsiciebiZdH6JRZK1s0qIE/OlzvAe35+A3gjwPNTgdQi9FEpS8UI5Nh4F6jLxgW8OPMtEo4f5tU2fXjzslvIcjgtbSMhzsmKYZ0svWaoIjWF0ArR+iWmK2NVqZKSlm5bkK9+9AAjZk3guo0LWH3G2dx+64tsqHmm5e04Y4Th3Rtbft1QRWoKoRWi9UtMA70qFTyZFLZM2xhD9/U/Mfz7iZQ/eYxXrr6TiS174opxBH9uAVmdJllYvimEKWnptBk9u8QH/mj9EtNAr6JeSlo6g/63kiwbhvI1D+/jpZnj6bjpF5bXacSgro+xuXq94E8shFCna8KdHhhtmSqRmAdfVLrxiIp6w6eutT7IG8Mtq2bx/XsPceXWFbzY7h5u7jvGtiAPkJGZRUpaer7neIJuekYmhr+DbrDnFUWwjVpU8dMRvYpK3qNaq8fxdQ7tZvS017lqaxqL6zXhmS6PsLVqeEaAwUbKxZEeGK2ZKtFEA72KOnbVqhGTQ58V0xny4weIMQzt+CCTmnfFSPj+MA4WtAsTdIs61ROtmSrRRAO9imiBgpDv8Xbn1Ti17D+mABtchyrpwC5GT3+dy7etYl79Zgzp+gg7Kp9haRuhyi9oFzToWjG/Hq2ZKtFEtxJUEcvfyNzpEGJjhMysnLD0ISbHxZ3Lv2XQ3I/JFgcj29/L5As7ggTaMrtg2pxdla37MguUDVQl3kl8mVi/I/CCbmUXaNvExIQ4FgxuH3KforE+TEmjWwmqEslvqQKXIcsVnsHJWft2MGbaOFqkr2f2WS14tvMA/qxU3dI2tu7LPBVQ/f2V4rspt9MhHDmefapcg+8IvKDpgVbNr0djpko00UCvIlZx3cxz5Li4d8kUnpg3iUxnWR6/7gmmNG5n2Sjem/dr9BcsW9SvelrQPnoiO8+G5L7z9gUJujq/XjpooFcRK1AQstO5e7YyZto4mu36jennXsY/Oz7EngpVbGsvWED1DdpnDv7O73mF/VLU+fXSQfPoVcQa1LkRcU7rV5f6E+vK5pEFn/Hth49R9+BuHu7+DA/0eNbWIC9Q4IBq9Z6kPZonMqpnUxIT4nT/2CimI3oVsbznm+0c2Tf+azPJqWO5YPfvfH3+1Yzo0J/98ZVtaw/+rjxZ0IBqxwhc59ejnwZ6FdE8QciOMgZlsrN4ZOF/eXDR/9gfX5n7eg5lVsPWll0/kKJsEBKttViUvTTQqxJh6R/7LQ3yzXZuZEzqOM7dt40vmlzDC9fcx6FyFSy7fiBjb21W5KCsI3BVUBroVcQbmrKaTxZts+RaZbNO8MT8Sdy7JIW/KlTlrpuH8+PZflOPLRfvjNEArYqFBnoVVvktrPH3GGBZkG+xYy1jUsdx1oGdTGrWhVFt/8GRsvGWXDsYZ4zwcs8Lw9KWUr400KuwyW+5PeD3MbGgJFn8yUwGzf2YO5d9S3rlmvS5dSQLGzQr8nWDSUyI03l0FRE00KuwCVRZccQ3a4kvE+v3saK67I+VvDLtNeoe3M1Hl3Qj+ao7OFbG/sVABS0hoJSdggZ6ESkHzAXKus//whgzzOecvsAz7l+PAA8aY1a6H9sKHAZcQHagWgwquqWkpQdMkTxwLOvUkn6rVDhxjGfnvE+fldPZUqUOvfqOZmnd8Gy/pwuOVKQJZUR/AmhvjDkiIk5gvohMM8Ys8jrnd+BqY8wBEekKTARaeT3ezhiz17puq5LEM2UTLm03L+XlGW9wxpH9vN2yJ/++oi8nnGXD0rZDRBccqYgTNNCb3PKWR9y/Ot3/jM85C71+XQTUtaqDKvIUtFKhvykbO1Q6foTnf3iHm9f8wK/Vknio3xBW1AnvyDrHGA3yKuKENEcvIg5gGXAOMN4Yszif0+8Bpnn9boCZImKAt40xEwO00R/oD5CUlBRKt1Qx8HdDddD/VjLim7VkHMvyG/jDUZys42+LeGnGeKoeO8hrl93KG5f35mSs05a2EuKceQqLeWgxMBWJQgr0xhgX0ExEEoApItLEGLPG9zwRaUduoL/C63AbY8xOEakJzBKRDcaYuX7amEjulA8tWrSIvCL5CghQOjjHBCybC/YWJ6ty7CAjvn+b7uvnsq7mmdx9y3DWnnG2LW15b849NGU1kxZtO+1PW52bV5GqQEXNjDEZwI9AF9/HRORC4F3gBmPMPq/n7HT/dzcwBWhZ+O6q4hbK6Nx7Y+iUtHSOnsi2viPGcN36ecx67yG6bFzIv67oS/c7XrUtyAMc9BrFj+zRlFdvbabFwFSJEErWTQ0gyxiTISJxQAfgFZ9zkoCvgNuNMb96HS8PxBhjDrt/7gS8YOULUOEV6ug8PSPT76jXCjWOHODFWW/S5defWVmrIX16P8qvNRpY3EpevtMyWopAlRShTN3UBj5yz9PHAJONMd+KyAMAxpgJwPNANeBNyd2cwZNGeQa5Uz2etj41xky3/mWocPFXPdEfwboVracYw41r5zDsh4nEZZ1gVNu7ePfSG3HF2F/KWKdlVEkWStbNKqC5n+MTvH6+F7jXzzlbgIuK2EcVQXyrJwYarVs9iq91aC8vzRzPNZuXsDTxfJ7u+ihbqoUvuatsrG7doEouXRmrCsx7yiLQ5tKWMYZbV83kudnv4cxxMeKa+/jo4m7khGEU7y0jMyvPTWalSgodpqgisXMXqLoH/+I/n/+TV6a/ztpaZ9P5H2/wQYsbbA/ybc6uisPP/rDeN5mVKkl0RK+KxN9GGPlN6YRCTA790lIZ/OOHGBGe6/QQnzbrghH7xyVlY2NYvu0gLuP/FRTXhuVKFYUGelUkKWnpjPhm7ak8+gNHTxQpyNc/sJMx016j1fY1zG3QnMFdH2FnpZrWdDaIOKeDsrExARdDgS6IUiWTBnpVaClp6Qz6YiVZrr9D+7GsnEJdKybHxd1Lp/LUvE/IcsQyqOuj/K9pB/AzhWIHz/Z+j32+It/z2p1XIyz9UcpKGuhVoSXP2HhakC+ss/duJ3naWC7euZFZ57TkuU4Ps7tiNQt6GJz3aleAJyevDDhtAzBnw55wdEspS2mgV4VW1GwbR46L+xd/yaMLPuWYM46B1z/F1POvDtsoPs7pYHj300sX5xfkQefoVcmkgV6FZGjKaj5bvD1oIAzVebt/Jzl1LE3/2sx3jdowrOMD7C1fxZJrhyIhzsnw7o3zpEomBln5q3P0qiTSQK/88pQiTs/IRLBuAZTTlcXDP0/m4Z8nk1GuIg/eMJhp510R/IkWK1821m8+fH4rf3V1rCqpNNCrPHxLEVsV5Jv8uYnk1LGcv2crUy5oy4gO/cmIq2TR1Qsm0BSMd7poekYmDhFcxpy6WauLpVRJpIFe5WH1RiFls08ycMFn3L/4S/aWT+Cem/7JD+e0Cv5EG+U3BaPFylS00UCvgNN3jbKyTs3F6esZkzqOc/bv4POmHXmp/T0cKlfBwhby54wRsnJOf0VOh+gUjCpVNNBHsVC3/POXD19U5bKO89Tc//CPpVPZWak6t/d6gXlnXmzZ9UPhmW4ZPnXtqUVQVeKdDLs+701YpaKZBvoo5W/Lv0BFuUZ8s9bSIN9q22pemfYaDTJ28Z/m1zL66rs4WjbesuuHyvPFpkFdlXYa6KOUv3l2T1Eu38DnKV9QVOVPHOOZnz7ijrTv+COhFr1ve5lFSRdacu2CSohzaoBXyk0DfZQKlFXiezwlLd2S9q74PY3R01+nzqE9vNfiBv7vytvJLFOuyNetEu8s8BeRv4VQSpVmWqY4SgXKKvE9/uxXq4rUTsUTRxk17TU+mfxPTsSW4ea+Y3jxmvssCfIAw65vjNMRfKWs7t2qVGA6oo9S/hb+eBb8WJVh027zEl6e/gY1jx7grVY3M/aKPpyILVP0zrvFyN/3E7wrZPpKTIhjweD2lrWrVLTRQB+lfOvEJ8Q7OZHlClqdMRSVMw8z7IeJ9Fw7hw3V63N/z+dYVfvcIl/XV59WSad+ji8Ty4FjWXlW6epqVaWCCxroRaQcMBco6z7/C2PMMJ9zBBgHXAscA+4yxix3P9bF/ZgDeNcYM9rSV6ACCmU0XFCdNy5k5Kw3Scg8zLjLb2P8Zb04Geu05Nre+rVOYmSPpn5X6XqCva5WVSo0oYzoTwDtjTFHRMQJzBeRacaYRV7ndAUauv+1At4CWomIAxgPdAR2AEtEZKoxZp2lr0LlkZKWznNTVnP0pDUrXKsdzWDE92/TbcM81pxxNnf0eoH1Nc+y5Nre4pyO0+bY/WUPeYK8TtcoFZqggd4YY4Aj7l+d7n++U7s3AB+7z10kIgkiUhtoAGwyxmwBEJH/us/VQG8D70JkljGG7uvnMvz7tyl/8hhjrrqDiS17ku2wftbPIZLnRmqo2UMeoS4SU6o0Cen/re6R+TLgHGC8MWaxzymJwHav33e4j/k7XrxFTqKU7xSHFWoc2c9LM9+k02+LWFH7XAZ1fZTfatS37PreBPhXr4vyBOU6AcoG+8sqKsgiMaVKk5DSK40xLmNMM6Au0FJEmvic4i//zeRzPA8R6S8iS0Vk6Z49uotPQVlaiMwYbl79Pd+/+yBX/b6cke3+Qc9+ybYFecj9UAQqGxzndJx2LNAN2PwWiSlVmhXo729jTIaI/Ah0AdZ4PbQDqOf1e11gJ1AmwHF/154ITARo0aKFlXW1SrRgUxG+m3MXVZ1Du3l5+nja/r6MxXUb80zXgWytav9oODFA3r9v9lB+0zEFnebx0OkeFe1CybqpAWS5g3wc0AF4xee0qcAA9xx8K+CgMWaXiOwBGorImUA60BvoY+kriGJDU1YzadG2U38CpWdk8vjnK1j6x35G9mjK0JTVfLJomzWNGUOfldMZMud9Yozh+Q7385+Lr8OI/WvqgqVIhlqvpiDTPB463aNKg1BG9LWBj9zz9DHAZGPMtyLyAIAxZgKQSm5q5SZy0yvvdj+WLSIDgBnkple+b4xZa/3LiD4paemnBXkPA0xyB3ergny9jD8ZPf012vyxigX1L+SZLgPZkVDLkmv7E+eMoWr5spaPoPNbJBZIQWoCKVVShZJ1swpo7uf4BK+fDfBwgOenkvtFoAogecbGgKtWDdYEeTE53LH8O5756UNcEsPgzgP470Wdbd+c+0R2ji3TJAWZ5vEo7HSPUiWJroyNUHYHmjP3pzMmdRyXpq9jzlmX8GznAeyqVMPWNj08+4B4pkmW/rGfORv2WBL8C1qWuDDTPUqVNBroI4D3zcCEeCfGWLdPq6+YHBf3LknhifmTOOFw8uS1j/Nlk/a2j+IDycxy5bkPEc458sJM9yhV0migL2a+NwOtyp7xp+GeP0ieNo5mu35lZsPWPNfpIfZUqGpbe6Hy/VIryBx5UTNmCjPdo1RJI7nT65GlRYsWZunSpcXdjbBoM3q2tStZ/Yh1ZXP/4i8ZuPAzjpSJZ1jHB/j2vCuLbRQfCgF+H31dvucEWiSWEOdkeHdrtgvU1EtVUojIMmNMC3+P6Yi+mNk9F3/BX1tITh1L491b+Oa8KxnW8QH2x1e2tU34Oy8+2JeYbzVKj1DmyAMtEsvIzLJk+kdTL1W00I1HiiAlLZ02o2dz5uDvaDN6dqF2a7Lrpl+Z7Cwen/cJX3/8ODWP7uf+G5/lkRueCUuQ98xx+1vV6owRqsQ7T20S0rd1UsgrX33l9yVpxYpYXWmrooWO6AupsKM978JjDhFcNkydXbRzI2OmjaPR3m182aQ9L7S/j4NxFS1vxx8R8hQmCzb10aJ+1UJNjwTKmPEo6l9LmnqpooUG+kIqzEIb3y8Hq4N82awTPD5/EvctSWF3+SrcdfMwfjz7UkvbCObVXs1Oe/2hpDsWNCXSw1/GjLei/rWkqZcqWujUTSEVZrRnaeExH5fsWEfqhwN54Jev+PzCjnS6982wBnkhd7OQcM5d92ieyKieTakSn3fjEytSJAtSUE2pSKYj+kIKdbRnddExX3EnjzNo7sfctewb0ivXpO+tI1nQoJktbQViZZZLQXn+GrAjO0ZTL1W00PTKQvKX2ue7O1JKWjqDvlhJlsue9/iyP1Yxevpr1M/4kw8v7saYq+/kWJnwTSvoVn5KRQ5Nr7RBKKO95BkbbQnyFU4cY/CPH9BvxTR+r1KbXn1G80s93y0C7OP7haaUimwa6Isg2E1EOxZCXbVlGaOmv0Htw3t559Ie/OvKfhx3lrO8HV8OEXKM0ekLpUogDfQ2KUxOfX4qHT/C0Nnv0mv19/xWrR439UsmLfE8S9sIREfwSpVsGugLKNSbflYuqunw22JemjmeakczeOOyXrx+eW9OxJax7PreBKgc50QEMo5l6QheqSiggb4Agi2SsjrDpsqxgwz7YSI91v3E+hoNuOem51lT6xxLru1PYkIcCwa3B07/QvN8aWmwV6pk0kBfAMGWxD/5v5W4cqy5+dp1w3xemDWByseP8O8r+vJW65vJcuTNF7eSJz9ca7woFV000BdAoJurOzMyGfHNWkuCfPWjB3hh5ltc++tCVtU6h369R7KxRoMiXxdy891F/JdCrhLvPC2TSLfXUyp6aKAPUUpaesBKiwYL6sgbQ491PzLs+4nEZx3nlavvZGLLnrhiHMGfG4KxtzY7Nb3kL/9/2PWNT/2uNV6Uii4a6EOU3x6uRXXG4b28NGM8HTYvYXmdRgzq+hibq9ez5Nq+GTOh5P9rjRelokvQQC8i9YCPgVpADjDRGDPO55xBQF+va54P1DDG7BeRrcBhwAVkB1q5FelsGc0awy2rZ/HP2e/hdGXzQvv7+PCSbuRYNIp3iPhNiwyW/6/b6ykVXUIZ0WcDTxpjlotIRWCZiMwyxqzznGCMSQaSAUTkeuBxY8x+r2u0M8bstbLj4RZfxsHRk9YVJEs8uJtR01/nqq1pLKrXhGe6DuSPKnUsu74zRki+5aJCzalrjRelokvQQG+M2QXscv98WETWA4nAugBPuQ34zLIeFhPv9MI4ZwzHsnIsua6YHPqumM7gHz9AjGFoxweZ1LwrRqwtJFqhXGyRAnNhSwcrpSJPgeboRaQB0BxYHODxeKALMMDrsAFmiogB3jbGTAzw3P5Af4CkpKSCdMtyvjcsrQrySQd28cr017hs22rmNmjOs10GsKPyGZZc21eGjZuMK6VKlpADvYhUAL4EHjPGHApw2vXAAp9pmzbGmJ0iUhOYJSIbjDFzfZ/o/gKYCLnVK0N+BRby3v3JSjE5Lu5a9i2D5n5MVoyDp7sMZPKFHW3dnFtvnCqlPEIK9CLiJDfITzLGfJXPqb3xmbYxxux0/3e3iEwBWgJ5An1x85d2aIWz921nTOo4Ltm5gR/OvpRnOz/MXxWrW9qGL71xqpTyFkrWjQDvAeuNMf/O57zKwNVAP69j5YEY99x+eaAT8EKRe10EgWrVWL37kyPHxX2/TOHx+ZPIdJblsW5PknJBW1tH8R5agEwp5S2UEX0b4HZgtYiscB97FkgCMMZMcB+7EZhpjDnq9dwzgCm53xXEAp8aY6Zb0O9CyW9pv5Xpk432bCU5dSwX/rmJaedezvMdH2RPhSqWXT8/iQlxGuSVUqcJJetmPrlFDYOd9yHwoc+xLcBFhexbkfgbuee3tD/QIqGCcLqyeHDRFwxY+DmHypXnoRsGk3reFUW6ZkEI6JSNUiqPqFwZG2jkHmhqZmdGJn1bJ/HJom2FbrPxn5v4v9SxnL9nK1+ffzXDO/TnQHzlQl/PnzinI+BrEKBvmDfnVkqVDFEZ6AON3B0iuPzskVvOGVPoIF8mO4uBCz/jgUVfsK98Avf1HMqshq0Lda1gRvVseuqvFK0Zr5QKVVQG+kDz7f6CPEBmIfPkm+3cSHLqWBru287kph0Y2f5eDpWrUKhrBeMQ0UVMSqlCicpAb8V8e37KZR3niXmTuGfp1/xZoRp33jKCn866xLb2IPCXlFJKBWPtuvsIMahzI+Kc1hQG83Xp9jVM++AR+i+Zwn8v6kTne8bbHuQhN5tGKaUKIypH9J7pjeFT15KRaU0pgPiTmTz900fctfxbtlU+g9t6v8TP9cOTUKQLoJRSRRGVgR44tQjKikDfZusKRk9/ncSDu/ngkusZc9WdZJYpZ0Ev/YtzxlDO6dAbrUopS0RtoIeiL4KqeOIoQ+a8T5+VM9hcNZFefUeztG7j4E8sAs9OUEopZZWoDvROh3DSVbibmG03L+XlGW9wxpH9TGjZk1ev6MsJZ1lL+hVoS0KllLJD1Ab6vu/8XKggXznzMM/Pfoeb1sxmY/UkHuwxhJV1ijY/nuCT897uvBpMWrTNb7B/cvJKHv98hU7ZKKUsE7WBfsHm/cFP8tHp158ZOfNNqh47yGuX3cobl/fmZKyz0H3w3a/VW6AFWp40Su86PBrslVJFEbWBviCqHjvIiFkTuH7DPNbWPIu7bxnO2jPOLtI1y5dx8NKNgatIJoaQ6++pw6OBXilVFFEX6IemrOazxdtDO9kYum2Yx4hZE6h44hj/d2U/JrS6mWxH6G+LI0aIweBZXCsCfVslMbJH03yf528Dbn9s2ZRcKVWqRFWgH5qyOuSaNTWO7GfkzDfp/NsiVtRuyNNdH+XXGg2CPs+q1EffDbhjAtTh0Z2ilFJFFVWBPqSRvDH0XDub5394h7isE7zc9m7eu7QHrpjAK2ljgC2jr7Ouo27etWv87XClC6WUUlaIqkAfrB5MrUN7eXnGG7TfspQliRfwTNeBbKlWN+h1/31rM4t6GJjvCF+zbpRSVomqQB+oDDHG0HvlDJ6d8z6xxsXwa/rz8cXXkZPPKF4g7MFWq1MqpewQVYH+tlb18szR1834k9HTX+eKP1ayMOlCnuk6kO0JtQJeI0bg3710dapSKnpEVaD3ZLp8tng7OTku7kxLZfDcj3AhPNv5YaZeei1Hs3JH/P5Wp8Y7Y3i554Ua5JVSUUVMkHltEakHfAzUAnKAicaYcT7ntAW+Bn53H/rKGPOC+7EuwDjAAbxrjBkdrFMtWrQwS5cuLdALOc1vv8E998C8edC5M0ycCElJhb+eUkpFOBFZZoxp4e+xUEb02cCTxpjlIlIRWCYis4wx63zOm2eM6ebTsAMYD3QEdgBLRGSqn+daw+WCsWNh6FAoVw4++ADuvDM3uV0ppUqpoIHeGLML2OX++bCIrAcSgVCCdUtgkzFmC4CI/Be4IcTnFsyBA9C1KyxeDN27w1tvQZ06ljejlFIlTYF2mBKRBkBzYLGfhy8TkZUiMk1EPLV8EwHv5PYd7mP+rt1fRJaKyNI9e/YUpFu5EhLgrLPg008hJUWDvFJKuYV8M1ZEKgBfAo8ZYw75PLwcqG+MOSIi1wIpQENy73n68ntTwBgzEZgIuXP0ofbLq4O5QV4ppdRpQhrRi4iT3CA/yRjzle/jxphDxpgj7p9TAaeIVCd3BF/P69S6wM4i91oppVTIggZ6ERHgPWC9MebfAc6p5T4PEWnpvu4+YAnQUETOFJEyQG9gqlWdV0opFVwoUzdtgNuB1SKywn3sWSAJwBgzAbgZeFBEsoFMoLfJzdvMFpEBwAxy0yvfN8astfYlKKWUyk/QPPriUOQ8eqWUKmXyy6MvUNaNUkqpkkcDvVJKRTkN9EopFeU00CulVJSLyJuxIrIH+KOQT68O7LWwO3bSvlqvpPQTtK92Ka19rW+MqeHvgYgM9EUhIksD3XmONNpX65WUfoL21S7a17x06kYppaKcBnqllIpy0RjoJxZ3BwpA+2q9ktJP0L7aRfvqI+rm6JVSSp0uGkf0SimlvGigV0qpKFciAr2I1BOROSKyXkTWisijfs5pKyIHRWSF+9/zXo91EZGNIrJJRAZHQF8HefVzjYi4RKSq+7GtIrLa/Zitld1EpJyI/OLeGWytiIzwc46IyGvu926ViFzs9Vg439dQ+trX3cdVIrJQRC7yeizS3tdI+byG0teI+Ly623OISJqIfOvnsYj4rIbY1/B+Vo0xEf8PqA1c7P65IvArcIHPOW2Bb/081wFsBs4CygArfZ8b7r76nH89MNvr961A9TC9rwJUcP/sJHeLyNY+51wLTHOf2xpYXEzvayh9vRyo4v65q6evEfq+RsrnNWhfI+Xz6m7vCeDTAO9dRHxWQ+xrWD+rJWJEb4zZZYxZ7v75MODZoDwUpzYoN8acBDwblNuiEH29DfjMrv7kx+Q64v7V6f7ne3f+BuBj97mLgAQRqU3439egfTXGLDTGHHD/uojcHc3CLsT3NZCIe199FNvnVUTqAtcB7wY4JSI+q6H0Ndyf1RIR6L2JjRuUWy1IXxGReKALuds0ehhgpogsE5H+YeijQ3I3lNkNzDLG+PY10PsX9vc1hL56u4fc0Z1HpL2vECGf11Df1wj4vI4FngZyAjweMZ9VgvfVm+2f1ZA3B48EYvMG5VYK0leP64EFxpj9XsfaGGN2ikhNYJaIbDDGzLWrn8YYF9BMRBKAKSLSxBizxuuUQO9f2N/XEPoKgIi0I/f/PFd4HY609zViPq+hvq8U4+dVRLoBu40xy0SkbaDT/BwL+2c1xL56zg3LZ7XEjOilBG1QHqyvXnrj82ewMWan+7+7gSnk/tlpO2NMBvAjuSM2b4Hev2Lb+D2fviIiF5L75/INxph9Xs+JqPc1kj6vwfrqpTg/r22A7iKyldypl/Yi8onPOZHyWQ2lr+H9rFo54W/XP3K/kT8GxuZzTi3+XgDWEtjmfl4ssAU4k79vxDQuzr66z6sM7AfKex0rD1T0+nkh0MXGvtYAEtw/xwHzgG4+51zH6Te4fnEfD/f7Gkpfk4BNwOU+xyPxfY2Uz2vQvkbK59Wr3bb4v8EZEZ/VEPsa1s9qSZm6KUkblIfSV4AbgZnGmKNezz2D3D+dIffD+akxZrqNfa0NfCQiDnL/uptsjPlWRB7w6msqudkMm4BjwN3ux8L9vobS1+eBasCb7vcw2+RWBozE9zVSPq+h9BUi4/OaR4R+VkPpa1g/q1oCQSmlolyJmaNXSilVOBrolVIqymmgV0qpKKeBXimlopwGeqWUinIa6JVSKsppoFdKqSj3/wLRaQu+Q7wFAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.scatter(y,yp)\n",
    "plt.plot([2.5,4.25],[2.5,4.25],color=\"red\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fc746846580>]"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#plt.scatter(diamonds[\"x\"],np.log10(diamonds[\"price\"]))\n",
    "plt.scatter(10**y,10**yp)\n",
    "plt.plot([0,17800],[0,17800],color=\"red\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(7605, 10)\n",
      "(32758, 10)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "42.89810697561005"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(diamonds[6.75<diamonds[\"x\"]].shape)\n",
    "print(diamonds[diamonds[\"x\"]<6.75].shape)\n",
    "sum((y-yp)**2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-77-9edb86a9bee0>:11: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  y_prediction[6.75<diamonds_test[\"x\"]]= 10**sel_func_2parts_1.predict(X1p)\n",
      "<ipython-input-77-9edb86a9bee0>:12: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  y_prediction[diamonds_test[\"x\"]<6.75]= 10**sel_func_2parts_2.predict(X2p)\n"
     ]
    }
   ],
   "source": [
    "diamonds_test= pd.read_csv(\"predict.csv\",index_col=\"id\")\n",
    "X_raw1p=cleaning_8(diamonds_test[6.75<diamonds_test[\"x\"]])\n",
    "X_raw2p=cleaning_8(diamonds_test[diamonds_test[\"x\"]<6.75])\n",
    "X1p=scaler_2parts.transform(X_raw1p)\n",
    "X2p=scaler_2parts.transform(X_raw2p)\n",
    "\n",
    "yp[6.75<diamonds[\"x\"]]=sel_func_2parts_1.predict(X1)\n",
    "yp[diamonds[\"x\"]<6.75]=sel_func_2parts_2.predict(X2)\n",
    "\n",
    "y_prediction = diamonds_test[\"x\"]\n",
    "y_prediction[6.75<diamonds_test[\"x\"]]= 10**sel_func_2parts_1.predict(X1p)\n",
    "y_prediction[diamonds_test[\"x\"]<6.75]= 10**sel_func_2parts_2.predict(X2p)\n",
    "\n",
    "diamonds_test[\"price\"]=y_prediction\n",
    "diamonds_test[\"price\"].to_csv(\"res/grid_2part_pred2.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id\n",
       "0        1035.622829\n",
       "1        6686.665156\n",
       "2         752.727315\n",
       "3        1827.579841\n",
       "4         953.213128\n",
       "            ...     \n",
       "13480     415.928165\n",
       "13481    4335.033166\n",
       "13482     907.950273\n",
       "13483     709.627372\n",
       "13484    4553.955654\n",
       "Name: price, Length: 13485, dtype: float64"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "diamonds_test[\"price\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}