20200707- final_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "RDKit WARNING: [16:58:47] Enabling RDKit 2019.09.3 jupyter extensions\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from rdkit import Chem\n",
    "from sklearn.model_selection import train_test_split,GridSearchCV, KFold\n",
    "from sklearn.linear_model import LinearRegression, Ridge, Lasso, LassoLarsCV, ElasticNetCV\n",
    "from sklearn.metrics import mean_squared_error\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.ensemble import RandomForestRegressor, ExtraTreesRegressor\n",
    "from sklearn.feature_selection import SelectKBest, mutual_info_regression, RFECV, SelectFromModel\n",
    "import pickle as pkl\n",
    "import seaborn as sns\n",
    "import xgboost as xgb\n",
    "from scipy.stats import uniform, randint\n",
    "import subprocess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def split_data(X_new):\n",
    "    X_final, validate = np.split(X_new.sample(frac=1), [int(.8*len(X_new))])\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X_final.drop(columns=['pChemBL'],inplace=False), X_final['pChemBL'], test_size=0.2, random_state=42)\n",
    "    X_validate = validate.drop(columns='pChemBL',inplace=False).copy()\n",
    "    y_validate = validate['pChemBL']\n",
    "    return X_train, X_test, X_validate, y_train, y_test, y_validate\n",
    "def train_RFR(X_train, X_test, X_validate, y_train, y_test, y_validate,estimators=150):\n",
    "    model2 = RandomForestRegressor(n_estimators=estimators)\n",
    "    model2.fit(X_train,y_train)\n",
    "    r_square = model2.score(X_train,y_train)\n",
    "    rmse = np.sqrt(mean_squared_error(model2.predict(X_train),y_train))\n",
    "    print('Number of estimators = {}'.format(estimators))\n",
    "    print(\"Training set results:\\nRMSE = {}\\t R^2 = {}\".format(rmse,r_square))\n",
    "    q_square = model2.score(X_test,y_test)\n",
    "    rmse = np.sqrt(mean_squared_error(model2.predict(X_test),y_test))\n",
    "    print(\"Test set results\\nRMSE = {}\\t Q^2 = {}\".format(rmse,q_square))\n",
    "    q_square = model2.score(X_validate,y_validate)\n",
    "    rmse = np.sqrt(mean_squared_error(y_validate,model2.predict(X_validate)))\n",
    "    print(\"Validation set results\\nRMSE = {}\\t Q^2 = {}\".format(rmse,q_square))    \n",
    "    return model2\n",
    "def split_train(X,estimators=150):\n",
    "    X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(X)\n",
    "    model = train_RFR(X_train, X_test, X_validate, y_train, y_test, y_validate,estimators)\n",
    "    return model\n",
    "def get_cols(n):\n",
    "    cols = []\n",
    "    for i in range(n):\n",
    "        cols.append('component'+str(i+1))\n",
    "    return cols\n",
    "def scores(y_test,predictions):\n",
    "    mse=mean_squared_error(predictions,y_test)\n",
    "    y_mean = np.full(shape=y_test.shape,fill_value=y_test.mean())\n",
    "    mse_ = mean_squared_error(y_mean,y_test)\n",
    "    print(\"RMSE = \",np.sqrt(mse))\n",
    "    print(\"Q^2 = \",1-(mse/mse_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "naAromAtom   1\n",
      "nAromBond   1\n",
      "BCUTc-1l   12\n",
      "BCUTc-1h   12\n",
      "BCUTp-1h   12\n",
      "SpMin1_Bhm   6\n",
      "SpMin2_Bhm   6\n",
      "SpMin3_Bhm   6\n",
      "SpMin4_Bhm   6\n",
      "SpMax1_Bhv   6\n",
      "SpMax2_Bhv   6\n",
      "SpMin1_Bhv   6\n",
      "SpMax1_Bhe   6\n",
      "SpMax3_Bhe   6\n",
      "SpMin1_Bhe   6\n",
      "SpMin2_Bhe   6\n",
      "SpMin4_Bhe   6\n",
      "SpMin6_Bhe   6\n",
      "SpMax1_Bhp   6\n",
      "SpMax2_Bhp   6\n",
      "SpMax3_Bhp   6\n",
      "SpMin1_Bhp   6\n",
      "SpMax1_Bhi   6\n",
      "SpMin1_Bhi   6\n",
      "SpMin2_Bhi   6\n",
      "SCH-6   2737\n",
      "VCH-5   2737\n",
      "SpMAD_Dt   332\n",
      "SHaaCH   10\n",
      "SaaN   10\n",
      "minHBa   10\n",
      "maxaaN   10\n",
      "hmin   10\n",
      "MDEC-33   2\n",
      "MDEN-22   2\n",
      "MDEN-23   2\n",
      "MLFER_A   1573\n",
      "MLFER_BH   1573\n",
      "piPC5   1\n",
      "piPC7   1\n",
      "piPC8   1\n",
      "piPC9   1\n",
      "piPC10   1\n",
      "TpiPC   1\n",
      "R_TpiPCTPC   1\n",
      "nTRing   4\n",
      "WTPT-4   796\n",
      "WTPT-5   796\n"
     ]
    }
   ],
   "source": [
    "xg2 = pd.read_csv('xgd2.csv')\n",
    "with open('./saved_models/xgreqd_cols.pkl','rb') as f:\n",
    "    cols = pkl.load(f)\n",
    "for col in cols:\n",
    "    count = xg2[col].isna().sum()\n",
    "    if count:\n",
    "        print(col,\" \",count)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "bad_cols = ['WTPT-4',\"WTPT-5\",\"SpMAD_Dt\",\"MLFER_A\",\"MLFER_BH\",\"SCH-6\",\"VCH-5\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "rfsk = pd.read_csv('./data/rfecvskb.csv')\n",
    "rfsk.drop(columns=bad_cols,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "reg = xgb.XGBRegressor()\n",
    "reg.load_model('./saved_models/xg_final.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(rfsk)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.7264129551692743\n",
      "Q^2 =  0.5841899518051463\n"
     ]
    }
   ],
   "source": [
    "reg.fit(X_train,y_train)\n",
    "predictions = reg.predict(X_test)\n",
    "scores(y_test,predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.7120686989419284\n",
      "Q^2 =  0.6195444768615059\n"
     ]
    }
   ],
   "source": [
    "predictions = reg.predict(X_validate)\n",
    "scores(y_validate,predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>38 descriptors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "second_chance = rfsk.drop(columns=[\"SHaaCH\",\"SaaN\",\"minHBa\",\"maxaaN\",\"hmin\"],inplace=False).copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 12 candidates, totalling 60 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  1.9min\n",
      "[Parallel(n_jobs=-1)]: Done  60 out of  60 | elapsed:  3.6min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5,\n",
       "             estimator=XGBRegressor(alpha=0.4, colsample_bytree=0.6,\n",
       "                                    learning_rate=0.01, max_depth=7,\n",
       "                                    min_child_weight=7, n_estimators=1000,\n",
       "                                    seed=27, subsample=0.8),\n",
       "             n_jobs=-1,\n",
       "             param_grid={'max_depth': range(5, 11, 2),\n",
       "                         'min_child_weight': [5, 7, 9, 11]},\n",
       "             scoring='neg_mean_squared_error', verbose=1)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(second_chance.drop(columns='pChemBL'),second_chance['pChemBL'],test_size=0.1,random_state=42)\n",
    "reg = xgb.XGBRegressor(alpha=0.4, colsample_bytree=0.6, max_depth=7, min_child_weight=7,\n",
    "             n_estimators=1000, seed=27, subsample=0.8,learning_rate=0.01)\n",
    "param_test = {\n",
    " 'max_depth':range(5,11,2),\n",
    " 'min_child_weight':[5,7,9,11]\n",
    "}\n",
    "gsearch1 = GridSearchCV(estimator = reg,param_grid = param_test, scoring='neg_mean_squared_error',n_jobs=-1, cv=5,verbose=1)\n",
    "gsearch1.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6559508772995691\n",
      "{'max_depth': 9, 'min_child_weight': 5}\n",
      "RMSE =  0.5974416428566585\n",
      "Q^2 =  0.7424214409982084\n"
     ]
    }
   ],
   "source": [
    "print(np.sqrt(0-gsearch1.best_score_))\n",
    "best38 = gsearch1.best_estimator_\n",
    "print(gsearch1.best_params_)\n",
    "preds = best38.predict(X_test)\n",
    "scores(y_test,preds)\n",
    "best38.save_model('./saved_models/best_from_gs38.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.22097044933258972\n",
      "Q^2 =  0.9639602803066302\n"
     ]
    }
   ],
   "source": [
    "preds = best38.predict(X_train)\n",
    "scores(y_train,preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.636264995258674\n",
      "Q^2 =  0.6873094497129433\n",
      "RMSE =  0.6191679058652316\n",
      "Q^2 =  0.749216606239191\n"
     ]
    }
   ],
   "source": [
    "reg = xgb.XGBRegressor(alpha=0.4, colsample_bytree=0.6,\n",
    "                                    learning_rate=0.01, max_depth=9,\n",
    "                                    min_child_weight=5, n_estimators=1000,\n",
    "                                    seed=27, subsample=0.8)\n",
    "X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(second_chance)\n",
    "reg.fit(X_train,y_train)\n",
    "predictions = reg.predict(X_test)\n",
    "scores(y_test,predictions)\n",
    "predictions = reg.predict(X_validate)\n",
    "scores(y_validate,predictions)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "reg.save_model('./saved_models/xg38.model')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>43 descriptors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fitting 5 folds for each of 12 candidates, totalling 60 fits\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[Parallel(n_jobs=-1)]: Using backend LokyBackend with 8 concurrent workers.\n",
      "[Parallel(n_jobs=-1)]: Done  34 tasks      | elapsed:  2.3min\n",
      "[Parallel(n_jobs=-1)]: Done  60 out of  60 | elapsed:  4.4min finished\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "GridSearchCV(cv=5,\n",
       "             estimator=XGBRegressor(alpha=0.4, colsample_bytree=0.6,\n",
       "                                    learning_rate=0.01, max_depth=7,\n",
       "                                    min_child_weight=7, n_estimators=1000,\n",
       "                                    seed=27, subsample=0.8),\n",
       "             n_jobs=-1,\n",
       "             param_grid={'max_depth': range(5, 11, 2),\n",
       "                         'min_child_weight': [5, 7, 9, 11]},\n",
       "             scoring='neg_mean_squared_error', verbose=1)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(rfsk.drop(columns='pChemBL'),rfsk['pChemBL'],test_size=0.1,random_state=42)\n",
    "reg = xgb.XGBRegressor(alpha=0.4, colsample_bytree=0.6, max_depth=7, min_child_weight=7,\n",
    "             n_estimators=1000, seed=27, subsample=0.8,learning_rate=0.01)\n",
    "param_test = {\n",
    " 'max_depth':range(5,11,2),\n",
    " 'min_child_weight':[5,7,9,11]\n",
    "}\n",
    "gsearch1 = GridSearchCV(estimator = reg, \n",
    " param_grid = param_test, scoring='neg_mean_squared_error',n_jobs=-1, cv=5,verbose=1)\n",
    "gsearch1.fit(X_train,y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.6491103160677638\n",
      "{'max_depth': 9, 'min_child_weight': 7}\n",
      "RMSE =  0.6107028774694965\n",
      "Q^2 =  0.7308597442202267\n"
     ]
    }
   ],
   "source": [
    "print(np.sqrt(0-gsearch1.best_score_))\n",
    "best = gsearch1.best_estimator_\n",
    "print(gsearch1.best_params_)\n",
    "preds = best.predict(X_test)\n",
    "scores(y_test,preds)\n",
    "best.save_model('./saved_models/best_from_gs.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "best.save_model('./saved_models/best43.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.6555279773133559\n",
      "Q^2 =  0.6747132687649459\n",
      "RMSE =  0.6160835524226539\n",
      "Q^2 =  0.7143639904362038\n"
     ]
    }
   ],
   "source": [
    "reg = xgb.XGBRegressor(alpha=0.4, colsample_bytree=0.6,\n",
    "                                    learning_rate=0.01, max_depth=9,\n",
    "                                    min_child_weight=7, n_estimators=1000,\n",
    "                                    seed=27, subsample=0.8)\n",
    "X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(rfsk)\n",
    "reg.fit(X_train,y_train)\n",
    "predictions = reg.predict(X_test)\n",
    "scores(y_test,predictions)\n",
    "predictions = reg.predict(X_validate)\n",
    "scores(y_validate,predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Verification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_padel(mol_folder_path,file_path,max_time='1500'):\n",
    "    Padel_path = 'C:\\\\Users\\\\HP\\\\PaDEL-Descriptor\\\\PaDEL-Descriptor.jar'\n",
    "    cmd_list = ['java','-jar',Padel_path, '-dir', mol_folder_path, '-2d','-file', file_path,'-maxruntime', max_time,\"-descriptortypes\", './saved_models/xg_desc3.xml','-usefilenameasmolname']\n",
    "    out = subprocess.Popen(cmd_list, \n",
    "           stdout=subprocess.PIPE, \n",
    "           stderr=subprocess.STDOUT)\n",
    "    stdout,stderr = out.communicate()\n",
    "    stdout = stdout.decode('utf-8')\n",
    "    with open('./Padel.txt','a') as f:\n",
    "        f.write(stdout)\n",
    "#get_padel(r'C:\\Users\\HP\\AZC_Internship\\jupyter notebooks\\mol','./xgdesc100.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "pChemBL\n"
     ]
    }
   ],
   "source": [
    "desc100 = pd.read_csv('./xgdesc100.csv')\n",
    "bad = []\n",
    "for col in desc100.columns:\n",
    "    if col not in second_chance.columns:\n",
    "        bad.append(col)\n",
    "for col in second_chance.columns:\n",
    "    if col not in desc100.columns:\n",
    "        print(col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "desc100.drop(columns=bad,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "naAromAtom    0\n",
       "nAromBond     0\n",
       "BCUTc-1l      3\n",
       "BCUTc-1h      3\n",
       "BCUTp-1h      3\n",
       "nBondsM       0\n",
       "SpMin1_Bhm    1\n",
       "SpMin2_Bhm    1\n",
       "SpMin3_Bhm    1\n",
       "SpMin4_Bhm    1\n",
       "SpMax1_Bhv    1\n",
       "SpMax2_Bhv    1\n",
       "SpMin1_Bhv    1\n",
       "SpMax1_Bhe    1\n",
       "SpMax3_Bhe    1\n",
       "SpMin1_Bhe    1\n",
       "SpMin2_Bhe    1\n",
       "SpMin4_Bhe    1\n",
       "SpMin6_Bhe    1\n",
       "SpMax1_Bhp    1\n",
       "SpMax2_Bhp    1\n",
       "SpMax3_Bhp    1\n",
       "SpMin1_Bhp    1\n",
       "SpMax1_Bhi    1\n",
       "SpMin1_Bhi    1\n",
       "SpMin2_Bhi    1\n",
       "nAtomP        0\n",
       "MDEC-33       1\n",
       "MDEN-22       1\n",
       "MDEN-23       1\n",
       "piPC5         0\n",
       "piPC7         0\n",
       "piPC8         0\n",
       "piPC9         0\n",
       "piPC10        0\n",
       "TpiPC         0\n",
       "R_TpiPCTPC    0\n",
       "nTRing        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "desc100.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>Re-evaluating those molecules</h2> - 100,10 and 11"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "naAromAtom    0\n",
       "nAromBond     0\n",
       "BCUTc-1l      0\n",
       "BCUTc-1h      0\n",
       "BCUTp-1h      0\n",
       "nBondsM       0\n",
       "SpMin1_Bhm    0\n",
       "SpMin2_Bhm    0\n",
       "SpMin3_Bhm    0\n",
       "SpMin4_Bhm    0\n",
       "SpMax1_Bhv    0\n",
       "SpMax2_Bhv    0\n",
       "SpMin1_Bhv    0\n",
       "SpMax1_Bhe    0\n",
       "SpMax3_Bhe    0\n",
       "SpMin1_Bhe    0\n",
       "SpMin2_Bhe    0\n",
       "SpMin4_Bhe    0\n",
       "SpMin6_Bhe    0\n",
       "SpMax1_Bhp    0\n",
       "SpMax2_Bhp    0\n",
       "SpMax3_Bhp    0\n",
       "SpMin1_Bhp    0\n",
       "SpMax1_Bhi    0\n",
       "SpMin1_Bhi    0\n",
       "SpMin2_Bhi    0\n",
       "nAtomP        0\n",
       "MDEC-33       0\n",
       "MDEN-22       0\n",
       "MDEN-23       0\n",
       "piPC5         0\n",
       "piPC7         0\n",
       "piPC8         0\n",
       "piPC9         0\n",
       "piPC10        0\n",
       "TpiPC         0\n",
       "R_TpiPCTPC    0\n",
       "nTRing        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "get_padel(r'C:\\Users\\HP\\AZC_Internship\\jupyter notebooks\\unevalmol','./xgdesc3.csv','-1')\n",
    "desc3 = pd.read_csv('./xgdesc3.csv')\n",
    "bad = []\n",
    "for col in desc3.columns:\n",
    "    if col not in second_chance.columns:\n",
    "        bad.append(col)\n",
    "desc3.drop(columns=bad,inplace=True)\n",
    "desc3.isna().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Conclusion:</b>Upon re-evaluation, sometimes it gives NaN, other times it works perfectly."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1> Using the models on X_main"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_main = pd.read_csv('./no_zeros_no_chembl.csv')\n",
    "y = X_main['pChemBL']\n",
    "bad = []\n",
    "for col in X_main.columns:\n",
    "    if col not in second_chance.columns:\n",
    "        bad.append(col)\n",
    "X_main.drop(columns=bad,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_main.drop(columns='pChemBL',inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Index(['naAromAtom', 'nAromBond', 'BCUTc-1l', 'BCUTc-1h', 'BCUTp-1h',\n",
      "       'nBondsM', 'SpMin1_Bhm', 'SpMin2_Bhm', 'SpMin3_Bhm', 'SpMin4_Bhm',\n",
      "       'SpMax1_Bhv', 'SpMax2_Bhv', 'SpMin1_Bhv', 'SpMax1_Bhe', 'SpMax3_Bhe',\n",
      "       'SpMin1_Bhe', 'SpMin2_Bhe', 'SpMin4_Bhe', 'SpMin6_Bhe', 'SpMax1_Bhp',\n",
      "       'SpMax2_Bhp', 'SpMax3_Bhp', 'SpMin1_Bhp', 'SpMax1_Bhi', 'SpMin1_Bhi',\n",
      "       'SpMin2_Bhi', 'nAtomP', 'MDEC-33', 'MDEN-22', 'MDEN-23', 'piPC5',\n",
      "       'piPC7', 'piPC8', 'piPC9', 'piPC10', 'TpiPC', 'R_TpiPCTPC', 'nTRing'],\n",
      "      dtype='object')\n"
     ]
    }
   ],
   "source": [
    "print(X_main.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "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>naAromAtom</th>\n",
       "      <th>nAromBond</th>\n",
       "      <th>BCUTc-1l</th>\n",
       "      <th>BCUTc-1h</th>\n",
       "      <th>BCUTp-1h</th>\n",
       "      <th>nBondsM</th>\n",
       "      <th>SpMin1_Bhm</th>\n",
       "      <th>SpMin2_Bhm</th>\n",
       "      <th>SpMin3_Bhm</th>\n",
       "      <th>SpMin4_Bhm</th>\n",
       "      <th>...</th>\n",
       "      <th>MDEN-22</th>\n",
       "      <th>MDEN-23</th>\n",
       "      <th>piPC5</th>\n",
       "      <th>piPC7</th>\n",
       "      <th>piPC8</th>\n",
       "      <th>piPC9</th>\n",
       "      <th>piPC10</th>\n",
       "      <th>TpiPC</th>\n",
       "      <th>R_TpiPCTPC</th>\n",
       "      <th>nTRing</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>16</td>\n",
       "      <td>18</td>\n",
       "      <td>-0.395197</td>\n",
       "      <td>0.211821</td>\n",
       "      <td>11.049269</td>\n",
       "      <td>18</td>\n",
       "      <td>2.082894</td>\n",
       "      <td>1.917433</td>\n",
       "      <td>1.844178</td>\n",
       "      <td>1.750754</td>\n",
       "      <td>...</td>\n",
       "      <td>1.190551</td>\n",
       "      <td>2.121320</td>\n",
       "      <td>6.437201</td>\n",
       "      <td>7.481230</td>\n",
       "      <td>7.991362</td>\n",
       "      <td>8.528420</td>\n",
       "      <td>8.939872</td>\n",
       "      <td>9.888853</td>\n",
       "      <td>11.822704</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>18</td>\n",
       "      <td>20</td>\n",
       "      <td>-0.324459</td>\n",
       "      <td>0.303911</td>\n",
       "      <td>11.858839</td>\n",
       "      <td>23</td>\n",
       "      <td>2.141224</td>\n",
       "      <td>2.023361</td>\n",
       "      <td>1.870068</td>\n",
       "      <td>1.784411</td>\n",
       "      <td>...</td>\n",
       "      <td>1.105209</td>\n",
       "      <td>1.190551</td>\n",
       "      <td>6.583841</td>\n",
       "      <td>7.714060</td>\n",
       "      <td>8.241999</td>\n",
       "      <td>8.756425</td>\n",
       "      <td>9.193246</td>\n",
       "      <td>10.122678</td>\n",
       "      <td>14.083920</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>27</td>\n",
       "      <td>29</td>\n",
       "      <td>-0.291902</td>\n",
       "      <td>0.416960</td>\n",
       "      <td>10.455854</td>\n",
       "      <td>30</td>\n",
       "      <td>2.049118</td>\n",
       "      <td>1.945557</td>\n",
       "      <td>1.911229</td>\n",
       "      <td>1.769681</td>\n",
       "      <td>...</td>\n",
       "      <td>2.576756</td>\n",
       "      <td>1.374730</td>\n",
       "      <td>6.481625</td>\n",
       "      <td>7.342774</td>\n",
       "      <td>7.783045</td>\n",
       "      <td>8.171347</td>\n",
       "      <td>8.584202</td>\n",
       "      <td>9.627425</td>\n",
       "      <td>11.948270</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>-0.386199</td>\n",
       "      <td>0.226178</td>\n",
       "      <td>11.054526</td>\n",
       "      <td>15</td>\n",
       "      <td>1.972555</td>\n",
       "      <td>1.908991</td>\n",
       "      <td>1.879489</td>\n",
       "      <td>1.851109</td>\n",
       "      <td>...</td>\n",
       "      <td>2.658167</td>\n",
       "      <td>0.927030</td>\n",
       "      <td>5.546397</td>\n",
       "      <td>5.544933</td>\n",
       "      <td>5.816739</td>\n",
       "      <td>5.858647</td>\n",
       "      <td>5.889570</td>\n",
       "      <td>7.770856</td>\n",
       "      <td>1.761710</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>21</td>\n",
       "      <td>23</td>\n",
       "      <td>-0.386171</td>\n",
       "      <td>0.198694</td>\n",
       "      <td>11.877160</td>\n",
       "      <td>25</td>\n",
       "      <td>2.019523</td>\n",
       "      <td>1.963038</td>\n",
       "      <td>1.877156</td>\n",
       "      <td>1.633266</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>6.608338</td>\n",
       "      <td>7.610075</td>\n",
       "      <td>8.146986</td>\n",
       "      <td>8.569907</td>\n",
       "      <td>8.994672</td>\n",
       "      <td>9.976050</td>\n",
       "      <td>14.000126</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 38 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   naAromAtom  nAromBond  BCUTc-1l  BCUTc-1h   BCUTp-1h  nBondsM  SpMin1_Bhm  \\\n",
       "0          16         18 -0.395197  0.211821  11.049269       18    2.082894   \n",
       "1          18         20 -0.324459  0.303911  11.858839       23    2.141224   \n",
       "2          27         29 -0.291902  0.416960  10.455854       30    2.049118   \n",
       "3           6          6 -0.386199  0.226178  11.054526       15    1.972555   \n",
       "4          21         23 -0.386171  0.198694  11.877160       25    2.019523   \n",
       "\n",
       "   SpMin2_Bhm  SpMin3_Bhm  SpMin4_Bhm  ...   MDEN-22   MDEN-23     piPC5  \\\n",
       "0    1.917433    1.844178    1.750754  ...  1.190551  2.121320  6.437201   \n",
       "1    2.023361    1.870068    1.784411  ...  1.105209  1.190551  6.583841   \n",
       "2    1.945557    1.911229    1.769681  ...  2.576756  1.374730  6.481625   \n",
       "3    1.908991    1.879489    1.851109  ...  2.658167  0.927030  5.546397   \n",
       "4    1.963038    1.877156    1.633266  ...  0.000000  0.000000  6.608338   \n",
       "\n",
       "      piPC7     piPC8     piPC9    piPC10      TpiPC  R_TpiPCTPC  nTRing  \n",
       "0  7.481230  7.991362  8.528420  8.939872   9.888853   11.822704       6  \n",
       "1  7.714060  8.241999  8.756425  9.193246  10.122678   14.083920       9  \n",
       "2  7.342774  7.783045  8.171347  8.584202   9.627425   11.948270       6  \n",
       "3  5.544933  5.816739  5.858647  5.889570   7.770856    1.761710       2  \n",
       "4  7.610075  8.146986  8.569907  8.994672   9.976050   14.000126       7  \n",
       "\n",
       "[5 rows x 38 columns]"
      ]
     },
     "execution_count": 128,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_main.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.28237514151854454\n",
      "Q^2 =  0.9412880184367602\n"
     ]
    }
   ],
   "source": [
    "preds=best38.predict(X_main)\n",
    "scores(y,preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins=np.linspace(4,10,7)\n",
    "plt.hist(reg.predict(X_main),alpha=0.5,label='preds',bins=bins)\n",
    "plt.hist(y,alpha=0.5,label='actual',bins=bins)\n",
    "plt.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "bins=np.linspace(4,10,7)\n",
    "plt.hist(preds,alpha=0.5,label='preds',bins=bins)\n",
    "plt.hist(y,alpha=0.5,label='actual',bins=bins)\n",
    "plt.legend(loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b> Both models perform well. In terms of statistics, best38 is better but by the looks of it, reg tends to predict the values outside the peak better"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h2>Scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [],
   "source": [
    "scaling = StandardScaler()\n",
    "scaling.fit(X_main)\n",
    "da = scaling.transform(X_main)\n",
    "X_scaled = pd.DataFrame(data=da,columns=X_main.columns)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.6678130296505691\n",
      "Q^2 =  0.6758148938639155\n",
      "RMSE =  0.641277694458705\n",
      "Q^2 =  0.690643713297785\n"
     ]
    }
   ],
   "source": [
    "X_scaled['pChemBL'] = y\n",
    "reg_scaled = xgb.XGBRegressor(alpha=0.4, colsample_bytree=0.6,\n",
    "                                    learning_rate=0.01, max_depth=9,\n",
    "                                    min_child_weight=7, n_estimators=1000,\n",
    "                                    seed=27, subsample=0.8)\n",
    "X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(X_scaled)\n",
    "reg_scaled.fit(X_train,y_train)\n",
    "predictions = reg_scaled.predict(X_test)\n",
    "scores(y_test,predictions)\n",
    "predictions = reg_scaled.predict(X_validate)\n",
    "scores(y_validate,predictions)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I tried twice, scaling offers <b>no major improvement</b>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Conclusion</b> Going to use best38(38 features without scaling)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Classification metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.read_csv('./no_zeros_no_chembl.csv')\n",
    "with open('./saved_models/good_columns','rb') as f:\n",
    "    cols = pkl.load(f)\n",
    "y = X['pChemBL']\n",
    "bad = []\n",
    "for col in X.columns:\n",
    "    if col not in cols:\n",
    "        bad.append(col)\n",
    "X.drop(columns=bad,inplace=True)\n",
    "reg = xgb.XGBRegressor()\n",
    "reg.load_model('./saved_models/best_from_gs38.model')\n",
    "preds = reg.predict(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[1422,   93],\n",
       "       [  84, 1262]], dtype=int64)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds = preds>7\n",
    "y = y>7\n",
    "from sklearn.metrics import confusion_matrix\n",
    "confusion_matrix(y,preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.934468715290633"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import f1_score\n",
    "f1_score(y,preds)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9313653136531366 0.937592867756315\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import precision_score, recall_score\n",
    "print(precision_score(y,preds),recall_score(y,preds))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<b>Conclusion:</b> Classification works pretty well"
   ]
  },
  {
   "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>nAcid</th>\n",
       "      <th>ALogP</th>\n",
       "      <th>ALogp2</th>\n",
       "      <th>AMR</th>\n",
       "      <th>apol</th>\n",
       "      <th>naAromAtom</th>\n",
       "      <th>nAromBond</th>\n",
       "      <th>nAtom</th>\n",
       "      <th>nHeavyAtom</th>\n",
       "      <th>nH</th>\n",
       "      <th>...</th>\n",
       "      <th>AMW</th>\n",
       "      <th>WTPT-1</th>\n",
       "      <th>WTPT-2</th>\n",
       "      <th>WTPT-3</th>\n",
       "      <th>WTPT-4</th>\n",
       "      <th>WTPT-5</th>\n",
       "      <th>WPATH</th>\n",
       "      <th>WPOL</th>\n",
       "      <th>XLogP</th>\n",
       "      <th>Zagreb</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>-0.7109</td>\n",
       "      <td>0.505379</td>\n",
       "      <td>70.4932</td>\n",
       "      <td>73.378583</td>\n",
       "      <td>16</td>\n",
       "      <td>18</td>\n",
       "      <td>65</td>\n",
       "      <td>34</td>\n",
       "      <td>31</td>\n",
       "      <td>...</td>\n",
       "      <td>7.157502</td>\n",
       "      <td>70.514766</td>\n",
       "      <td>2.073964</td>\n",
       "      <td>27.938805</td>\n",
       "      <td>11.267053</td>\n",
       "      <td>16.671753</td>\n",
       "      <td>3404.0</td>\n",
       "      <td>59.0</td>\n",
       "      <td>2.188</td>\n",
       "      <td>184.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>-1.4578</td>\n",
       "      <td>2.125181</td>\n",
       "      <td>44.0418</td>\n",
       "      <td>64.304653</td>\n",
       "      <td>18</td>\n",
       "      <td>20</td>\n",
       "      <td>52</td>\n",
       "      <td>31</td>\n",
       "      <td>21</td>\n",
       "      <td>...</td>\n",
       "      <td>7.830281</td>\n",
       "      <td>65.167083</td>\n",
       "      <td>2.102164</td>\n",
       "      <td>17.955201</td>\n",
       "      <td>2.565237</td>\n",
       "      <td>15.389964</td>\n",
       "      <td>2593.0</td>\n",
       "      <td>55.0</td>\n",
       "      <td>6.575</td>\n",
       "      <td>180.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0.0748</td>\n",
       "      <td>0.005595</td>\n",
       "      <td>27.2874</td>\n",
       "      <td>69.268860</td>\n",
       "      <td>27</td>\n",
       "      <td>29</td>\n",
       "      <td>57</td>\n",
       "      <td>37</td>\n",
       "      <td>20</td>\n",
       "      <td>...</td>\n",
       "      <td>8.827511</td>\n",
       "      <td>76.232497</td>\n",
       "      <td>2.060338</td>\n",
       "      <td>31.795843</td>\n",
       "      <td>2.556014</td>\n",
       "      <td>22.098211</td>\n",
       "      <td>4979.0</td>\n",
       "      <td>56.0</td>\n",
       "      <td>5.589</td>\n",
       "      <td>198.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>-10.0458</td>\n",
       "      <td>100.918098</td>\n",
       "      <td>183.2952</td>\n",
       "      <td>142.418303</td>\n",
       "      <td>6</td>\n",
       "      <td>6</td>\n",
       "      <td>135</td>\n",
       "      <td>64</td>\n",
       "      <td>71</td>\n",
       "      <td>...</td>\n",
       "      <td>6.678230</td>\n",
       "      <td>125.580825</td>\n",
       "      <td>1.962200</td>\n",
       "      <td>61.611922</td>\n",
       "      <td>20.276064</td>\n",
       "      <td>41.335857</td>\n",
       "      <td>21440.0</td>\n",
       "      <td>92.0</td>\n",
       "      <td>-0.194</td>\n",
       "      <td>302.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>-1.7897</td>\n",
       "      <td>3.203026</td>\n",
       "      <td>32.3564</td>\n",
       "      <td>58.685481</td>\n",
       "      <td>21</td>\n",
       "      <td>23</td>\n",
       "      <td>47</td>\n",
       "      <td>30</td>\n",
       "      <td>17</td>\n",
       "      <td>...</td>\n",
       "      <td>8.491716</td>\n",
       "      <td>62.402934</td>\n",
       "      <td>2.080098</td>\n",
       "      <td>16.872064</td>\n",
       "      <td>13.655728</td>\n",
       "      <td>3.216336</td>\n",
       "      <td>2319.0</td>\n",
       "      <td>52.0</td>\n",
       "      <td>4.844</td>\n",
       "      <td>168.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 1444 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   nAcid    ALogP      ALogp2       AMR        apol  naAromAtom  nAromBond  \\\n",
       "0      0  -0.7109    0.505379   70.4932   73.378583          16         18   \n",
       "1      0  -1.4578    2.125181   44.0418   64.304653          18         20   \n",
       "2      0   0.0748    0.005595   27.2874   69.268860          27         29   \n",
       "3      0 -10.0458  100.918098  183.2952  142.418303           6          6   \n",
       "4      0  -1.7897    3.203026   32.3564   58.685481          21         23   \n",
       "\n",
       "   nAtom  nHeavyAtom  nH  ...       AMW      WTPT-1    WTPT-2     WTPT-3  \\\n",
       "0     65          34  31  ...  7.157502   70.514766  2.073964  27.938805   \n",
       "1     52          31  21  ...  7.830281   65.167083  2.102164  17.955201   \n",
       "2     57          37  20  ...  8.827511   76.232497  2.060338  31.795843   \n",
       "3    135          64  71  ...  6.678230  125.580825  1.962200  61.611922   \n",
       "4     47          30  17  ...  8.491716   62.402934  2.080098  16.872064   \n",
       "\n",
       "      WTPT-4     WTPT-5    WPATH  WPOL  XLogP  Zagreb  \n",
       "0  11.267053  16.671753   3404.0  59.0  2.188   184.0  \n",
       "1   2.565237  15.389964   2593.0  55.0  6.575   180.0  \n",
       "2   2.556014  22.098211   4979.0  56.0  5.589   198.0  \n",
       "3  20.276064  41.335857  21440.0  92.0 -0.194   302.0  \n",
       "4  13.655728   3.216336   2319.0  52.0  4.844   168.0  \n",
       "\n",
       "[5 rows x 1444 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df= pd.read_csv('C:/Users/HP/AZC_Internship/Data/Descriptor data/Padel descriptors/actual_padel_AKT.csv')\n",
    "df.drop(columns='Name').head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = xgb.XGBRegressor()\n",
    "model.load_model('./saved_models/best_from_gs38.model')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pd.read_csv('./no_zeros_no_chembl.csv')\n",
    "with open('./saved_models/good_columns','rb') as f:\n",
    "    cols = pkl.load(f)\n",
    "y = X['pChemBL']\n",
    "bad = []\n",
    "for col in X.columns:\n",
    "    if col not in cols:\n",
    "        bad.append(col)\n",
    "X.drop(columns=bad,inplace=True)\n",
    "X['pChemBL'] = y\n",
    "X_train, X_test, X_validate, y_train, y_test, y_validate = split_data(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test set results\n",
      "RMSE =  0.2658655187652451\n",
      "Q^2 =  0.9469764088333621\n",
      "Validation set results\n",
      "RMSE =  0.2939870719281471\n",
      "Q^2 =  0.9404091505555041\n"
     ]
    }
   ],
   "source": [
    "#model.fit(X_train,y_train)\n",
    "print(\"Test set results\")\n",
    "scores(y_test,model.predict(X_test))\n",
    "print(\"Validation set results\")\n",
    "scores(y_validate,model.predict(X_validate))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 1080x1080 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xgb.plot_importance(model,max_num_features=10)\n",
    "plt.figure(figsize=(15,15))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<BarContainer object of 5 artists>"
      ]
     },
     "execution_count": 29,
     "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": [
    "q2 = [0.46,.38,0.57,0.56,0.73]\n",
    "x = ['PCA_Lasso','AE_Linear','Corr_RFR','PCA_RFR','XGB']\n",
    "plt.ylabel('Q^2')\n",
    "plt.xlabel('Models')\n",
    "plt.bar(x,q2,width=0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "graph = xgb.to_graphviz(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "digraph {\n",
      "\tgraph [rankdir=UT]\n",
      "\t0 [label=\"SpMax1_Bhe<3.95549011\" shape=circle]\n",
      "\t0 -> 1 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t0 -> 2 [label=no color=\"#FF0000\"]\n",
      "\t1 [label=\"SpMin2_Bhe<1.8454138\" shape=circle]\n",
      "\t1 -> 3 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t1 -> 4 [label=no color=\"#FF0000\"]\n",
      "\t3 [label=\"nAtomP<7\" shape=circle]\n",
      "\t3 -> 7 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t3 -> 8 [label=no color=\"#FF0000\"]\n",
      "\t7 [label=\"SpMax3_Bhp<3.60549307\" shape=circle]\n",
      "\t7 -> 13 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t7 -> 14 [label=no color=\"#FF0000\"]\n",
      "\t13 [label=\"leaf=0.0448095202\" shape=box]\n",
      "\t14 [label=\"leaf=0.0239940397\" shape=box]\n",
      "\t8 [label=\"leaf=0.0598164275\" shape=box]\n",
      "\t4 [label=\"leaf=0.036032822\" shape=box]\n",
      "\t2 [label=\"SpMax1_Bhp<4.04902792\" shape=circle]\n",
      "\t2 -> 5 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t2 -> 6 [label=no color=\"#FF0000\"]\n",
      "\t5 [label=\"SpMin6_Bhe<0.991659462\" shape=circle]\n",
      "\t5 -> 9 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t5 -> 10 [label=no color=\"#FF0000\"]\n",
      "\t9 [label=\"leaf=0.0545261353\" shape=box]\n",
      "\t10 [label=\"SpMin3_Bhm<1.94925642\" shape=circle]\n",
      "\t10 -> 15 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t10 -> 16 [label=no color=\"#FF0000\"]\n",
      "\t15 [label=\"SpMin2_Bhe<1.74087763\" shape=circle]\n",
      "\t15 -> 19 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t15 -> 20 [label=no color=\"#FF0000\"]\n",
      "\t19 [label=\"leaf=0.0619191863\" shape=box]\n",
      "\t20 [label=\"SpMin6_Bhe<1.28945851\" shape=circle]\n",
      "\t20 -> 25 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t20 -> 26 [label=no color=\"#FF0000\"]\n",
      "\t25 [label=\"SpMin2_Bhe<1.84156573\" shape=circle]\n",
      "\t25 -> 29 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t25 -> 30 [label=no color=\"#FF0000\"]\n",
      "\t29 [label=\"leaf=0.0688814372\" shape=box]\n",
      "\t30 [label=\"leaf=0.0612301342\" shape=box]\n",
      "\t26 [label=\"leaf=0.0720620751\" shape=box]\n",
      "\t16 [label=\"leaf=0.0438333303\" shape=box]\n",
      "\t6 [label=\"SpMax1_Bhe<4.19803333\" shape=circle]\n",
      "\t6 -> 11 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t6 -> 12 [label=no color=\"#FF0000\"]\n",
      "\t11 [label=\"MDEN-23<0.974414825\" shape=circle]\n",
      "\t11 -> 17 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t11 -> 18 [label=no color=\"#FF0000\"]\n",
      "\t17 [label=\"piPC7<6.24557066\" shape=circle]\n",
      "\t17 -> 21 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t17 -> 22 [label=no color=\"#FF0000\"]\n",
      "\t21 [label=\"leaf=0.0354285724\" shape=box]\n",
      "\t22 [label=\"leaf=0.0633456632\" shape=box]\n",
      "\t18 [label=\"MDEN-22<0.224999994\" shape=circle]\n",
      "\t18 -> 23 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t18 -> 24 [label=no color=\"#FF0000\"]\n",
      "\t23 [label=\"leaf=0.0485799983\" shape=box]\n",
      "\t24 [label=\"MDEN-22<1.81126571\" shape=circle]\n",
      "\t24 -> 27 [label=\"yes, missing\" color=\"#0000FF\"]\n",
      "\t24 -> 28 [label=no color=\"#FF0000\"]\n",
      "\t27 [label=\"leaf=0.0617084019\" shape=box]\n",
      "\t28 [label=\"leaf=0.0547258779\" shape=box]\n",
      "\t12 [label=\"leaf=0.0740183666\" shape=box]\n",
      "}\n"
     ]
    }
   ],
   "source": [
    "print(graph)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<h1>Demonstration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>Actual</th>\n",
       "      <th>Predicted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>6.82</td>\n",
       "      <td>6.752529</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7.64</td>\n",
       "      <td>7.553743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>4.30</td>\n",
       "      <td>4.454226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>6.03</td>\n",
       "      <td>6.142269</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6.42</td>\n",
       "      <td>6.426538</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   Actual  Predicted\n",
       "0    6.82   6.752529\n",
       "1    7.64   7.553743\n",
       "2    4.30   4.454226\n",
       "3    6.03   6.142269\n",
       "4    6.42   6.426538"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('./no_zeros_no_chembl.csv')\n",
    "y = df['pChemBL']\n",
    "with open('./saved_models/good_columns','rb') as f:\n",
    "    cols = pkl.load(f)\n",
    "bad = []\n",
    "for col in df.columns:\n",
    "    if col not in cols:\n",
    "        bad.append(col\n",
    "df.drop(columns=bad,inplace=True)\n",
    "regressor = xgb.XGBRegressor()\n",
    "regressor.load_model('./saved_models/best_from_gs38.model')\n",
    "X = pd.DataFrame(columns=['Actual','Predicted'])\n",
    "preds = regressor.predict(df)\n",
    "X['Actual'] = y\n",
    "X['Predicted'] = preds\n",
    "X.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "RMSE =  0.28237514151854454\n",
      "Q^2 =  0.9412880184367602\n"
     ]
    }
   ],
   "source": [
    "scores(y,preds)"
   ]
  },
  {
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}