Yield Prediction using Molecular Descriptors_GetAvailableProperties_ Condensation_reactions.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "8ef8447e-1e4b-4962-8982-7a2102315fd8",
   "metadata": {},
   "source": [
    "# Yield Prediction using Molecular Descriptors_GetAvailablePropertiers_ Condensation_reactions"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3fa5cf84-8ed4-4846-b913-5440b27700fe",
   "metadata": {},
   "source": [
    "Found this dataset from a recent publication by a D.M. Makarov et al. (Journal of Computational Science 74 (2023) 102173) (https://doi.org/10.1016/j.jocs.2023.102173)\n",
    "\n",
    "\"\"\"We considered pyrrole or dipyrromethane condensation reactions with various aldehydes, resulting in the production of boron(III) dipyrromethene or BODIPY (681 records). These reactions were retrieved from articles (see “Dataset reactions” and Scheme S1). Addi tionally, we used the reactions of the production of dipyrromethane (111 records) and porphyrins (457 records). All condensation reactions for dipyrromethanes and 213 reactions for porphyrins with various al dehydes were obtained in our laboratory. The remaining 244 reactions for the porphyrins synthesis were obtained from articles (see “Dataset reactions”). Our experimental dataset is based on a study of pyrrole condensation processes with aldehydes, using catalytic amounts of organic acids to produce ms-aryl- and ß-alkyl-substituted dipyrro methanes. \"\"\"\n",
    "\n",
    "The objective of this notebook is to introduce a method for predicting yield\n",
    "\n",
    "As a beginner, there may be numerous opportunities for improvement in this notebook. I was largely inspired by the work of D.M. Makarov et al and the STEPHEN LEE's notebook (BELKA: Molecule Representations for ML Tutorial) thanks to them."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "371b98fd-d27f-4e26-bb7a-9efaecc8a32a",
   "metadata": {},
   "source": [
    "## Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b142db6d-8e35-4799-adc5-1c563a11d05f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "64d1a451-98a8-4f2a-a1e8-12513aacd312",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.max_row', 211)\n",
    "pd.set_option('display.max_columns',211)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "cca43b9b-1699-4134-843a-0738f1af67ca",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1249, 4)\n"
     ]
    },
    {
     "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>SMILES</th>\n",
       "      <th>yield</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Ind</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...</td>\n",
       "      <td>0.05</td>\n",
       "      <td>20</td>\n",
       "      <td>78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...</td>\n",
       "      <td>1.00</td>\n",
       "      <td>40</td>\n",
       "      <td>909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...</td>\n",
       "      <td>2.00</td>\n",
       "      <td>20</td>\n",
       "      <td>747</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...</td>\n",
       "      <td>2.10</td>\n",
       "      <td>20</td>\n",
       "      <td>77</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...</td>\n",
       "      <td>2.50</td>\n",
       "      <td>25</td>\n",
       "      <td>266</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              SMILES  yield  Temperature  Ind\n",
       "0  CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...   0.05           20   78\n",
       "1  O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...   1.00           40  909\n",
       "2  c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...   2.00           20  747\n",
       "3  CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...   2.10           20   77\n",
       "4  Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...   2.50           25  266"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Téléchargement du Dataset \n",
    "DataFrame = r\"C:\\Users\\loris\\Desktop\\IA_chemistry\\Dipyrromethanes_condensation_reactions-main\\Condensation_reactions - Copie.xlsx\"\n",
    "\n",
    "df = pd.read_excel(DataFrame)\n",
    "\n",
    "df= df.copy()\n",
    "\n",
    "print(df.shape)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "67ffd110-c6d9-455a-ad96-bb39af5fd3c4",
   "metadata": {},
   "source": [
    "## Missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "5f14037c-4205-451c-960f-ddc6b8af5bad",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SMILES         0\n",
       "yield          0\n",
       "Temperature    0\n",
       "Ind            0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Recherche des valeurs manquantes\n",
    "missing_values = df.isnull().sum()\n",
    "missing_values"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9cb3159e-11e9-4390-bd10-e6f2258c9c1b",
   "metadata": {},
   "source": [
    "## target "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b0862b97-d681-4219-a66c-36081d285e48",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Création de la figure et des axes pour les subplots\n",
    "fig, axs = plt.subplots(2, 1, figsize=(10,8))\n",
    "\n",
    "# Tracé de l'histogramme sur le premier subplot\n",
    "sns.histplot(data=df, x='yield', bins=30, ax=axs[0])\n",
    "axs[0].set_xlabel('Rendement chimique (%)')\n",
    "axs[0].set_ylabel(\"Nombre d'occurences\")\n",
    "axs[0].set_title(\"Distribution des rendements chimiques\")\n",
    "\n",
    "# Tracé du boxplot sur le deuxième subplot\n",
    "sns.boxplot(df[\"yield\"], ax=axs[1])\n",
    "axs[1].set_ylabel(\"Rendement chimique (%)\")\n",
    "axs[1].set_title(\"Distribution des rendements chimiques\")\n",
    "\n",
    "# Ajustement de l'espacement entre les subplots\n",
    "plt.tight_layout()\n",
    "\n",
    "# Affichage des graphiques \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f7dcd422-f8bb-4167-8d5e-35823cc05676",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d9033592-59e3-49c2-baf1-bd5e62937e46",
   "metadata": {},
   "source": [
    "## Obtenir des Mols"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "02f237b7-52d5-4ec6-845d-a2278927052f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from rdkit import Chem\n",
    "from rdkit.Chem import AllChem"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "09a3e1c4-d665-44e8-b225-4c6f066e2d21",
   "metadata": {},
   "outputs": [],
   "source": [
    "def change_smile(df, columns_names):\n",
    "    df[\"SMILES_2\"] = df[columns_names].replace('>>','.', regex=True)\n",
    "    return df "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "4c5184b9-b0e4-4392-9099-30a0f247e752",
   "metadata": {},
   "outputs": [],
   "source": [
    "def Mols(df, columns_names):\n",
    "    df['mols'] = pd.Series(df[columns_names]).apply(Chem.MolFromSmiles)\n",
    "    return df "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "a4fd21cd-e92e-4f94-8acb-0d03dd1a0468",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = change_smile(df, \"SMILES\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "6c683e57-2676-4622-8f29-f089a958392d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[17:26:53] Explicit valence for atom # 26 H, 2, is greater than permitted\n",
      "[17:26:54] Explicit valence for atom # 26 H, 2, is greater than permitted\n",
      "[17:26:54] Explicit valence for atom # 0 O, 3, is greater than permitted\n",
      "[17:26:54] Explicit valence for atom # 10 O, 3, is greater than permitted\n"
     ]
    }
   ],
   "source": [
    "df = Mols(df, \"SMILES_2\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "478849e8-accb-462a-ab87-293ee5889f98",
   "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>SMILES</th>\n",
       "      <th>yield</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>Ind</th>\n",
       "      <th>SMILES_2</th>\n",
       "      <th>mols</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...</td>\n",
       "      <td>0.05</td>\n",
       "      <td>20</td>\n",
       "      <td>78</td>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...</td>\n",
       "      <td>&lt;rdkit.Chem.rdchem.Mol object at 0x000002096B1...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...</td>\n",
       "      <td>1.00</td>\n",
       "      <td>40</td>\n",
       "      <td>909</td>\n",
       "      <td>O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...</td>\n",
       "      <td>&lt;rdkit.Chem.rdchem.Mol object at 0x000002096B1...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...</td>\n",
       "      <td>2.00</td>\n",
       "      <td>20</td>\n",
       "      <td>747</td>\n",
       "      <td>c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...</td>\n",
       "      <td>&lt;rdkit.Chem.rdchem.Mol object at 0x000002096B1...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...</td>\n",
       "      <td>2.10</td>\n",
       "      <td>20</td>\n",
       "      <td>77</td>\n",
       "      <td>CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...</td>\n",
       "      <td>&lt;rdkit.Chem.rdchem.Mol object at 0x000002096B1...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...</td>\n",
       "      <td>2.50</td>\n",
       "      <td>25</td>\n",
       "      <td>266</td>\n",
       "      <td>Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...</td>\n",
       "      <td>&lt;rdkit.Chem.rdchem.Mol object at 0x000002096B1...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                              SMILES  yield  Temperature  Ind  \\\n",
       "0  CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...   0.05           20   78   \n",
       "1  O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...   1.00           40  909   \n",
       "2  c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...   2.00           20  747   \n",
       "3  CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...   2.10           20   77   \n",
       "4  Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...   2.50           25  266   \n",
       "\n",
       "                                            SMILES_2  \\\n",
       "0  CC(C)(C)c1ccc(OCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C=...   \n",
       "1  O=Cc1ccccc1OCCOC(=O)C#CC(=O)OCCOc1ccccc1C=O.c1...   \n",
       "2  c1cc[nH]c1.Cc1cc(C)c(C=O)c(C)c1.COC(=O)c1ccc(C...   \n",
       "3  CC(C)(C)c1ccc(OCCCCCOc2ccc(C(C)(C)C)cc2C=O)c(C...   \n",
       "4  Cc1ccc(-c2ccccc2C=O)cc1.Fc1c(F)c(F)c(C(c2ccc[n...   \n",
       "\n",
       "                                                mols  \n",
       "0  <rdkit.Chem.rdchem.Mol object at 0x000002096B1...  \n",
       "1  <rdkit.Chem.rdchem.Mol object at 0x000002096B1...  \n",
       "2  <rdkit.Chem.rdchem.Mol object at 0x000002096B1...  \n",
       "3  <rdkit.Chem.rdchem.Mol object at 0x000002096B1...  \n",
       "4  <rdkit.Chem.rdchem.Mol object at 0x000002096B1...  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4d630bb6-ed61-4bb9-a2f5-f76addb24624",
   "metadata": {},
   "source": [
    "### Enlève les dernières valeurs manquantes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "94bd887d-7d17-4e3a-a77c-08cd1ebe8a22",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SMILES         0\n",
       "yield          0\n",
       "Temperature    0\n",
       "Ind            0\n",
       "SMILES_2       0\n",
       "mols           4\n",
       "dtype: int64"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "96294bcf-774b-461a-8c67-b8d8df6806e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.dropna()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "1c19fc21-9a13-464e-9845-3ae7f805b976",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SMILES         0\n",
       "yield          0\n",
       "Temperature    0\n",
       "Ind            0\n",
       "SMILES_2       0\n",
       "mols           0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d27374f9-8b11-45e1-8b48-24820aa04b51",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1245, 6)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "06b9e0d7-7a28-4d4e-88e3-21165659a78e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "3bd6d59f-a284-42db-8972-7f7d348ce363",
   "metadata": {},
   "source": [
    "## 🧪 Molecular Descriptors"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90dd3376-4f51-4278-90ac-f699dbdf20e5",
   "metadata": {},
   "source": [
    "Molecular descriptors are quantitative properties that capture information about the chemical structure that can be used effectively in cheminformatics, particularly for the modeling and analysis of chemical compounds. Descriptors can pertain to various aspects of a chemical structure, such as its constitution, topology, geometry, electron distribution, and hydrophobicity.\n",
    "\n",
    "The following code demonstrates how to generate molecular descriptors from SMILES."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "6f8ce448-98bf-4bf6-b0a9-a857ae99b866",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "import pandas as pd\n",
    "from rdkit.Chem import Descriptors\n",
    "from rdkit.ML.Descriptors import MoleculeDescriptors"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "38a5f86e-5ff4-4a0e-aeb7-0686b9ccbbe0",
   "metadata": {},
   "outputs": [],
   "source": [
    "def calculate_descriptors(df):\n",
    "    descriptor_names = list(Chem.rdMolDescriptors.Properties.GetAvailableProperties())\n",
    "    descriptor_names.remove('exactmw')\n",
    "    for name in tqdm(descriptor_names):\n",
    "        df[name] = df['SMILES_2'].apply(lambda x: descriptor(name, x))\n",
    "    return df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "dc2151b6-1294-4cf5-95e9-d06aa4b1b831",
   "metadata": {},
   "outputs": [],
   "source": [
    "def descriptor(descriptor:str, smi:str)->float:\n",
    "    get_descriptors = Chem.rdMolDescriptors.Properties([descriptor])\n",
    "    try:\n",
    "        return get_descriptors.ComputeProperties(Chem.MolFromSmiles(smi))[0]\n",
    "    except:\n",
    "        return None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "34207289-b26a-434d-9fed-2570a7a02b25",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 42/42 [00:37<00:00,  1.11it/s]\n"
     ]
    }
   ],
   "source": [
    "df = calculate_descriptors(df)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "88bb3475-cf69-400f-8a9e-5dea7c1ff8b8",
   "metadata": {},
   "outputs": [],
   "source": [
    "df_2 = df.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "9e2b8b79-6255-4262-8a10-01c6850306b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Sauvegarde du DataFrame dans un fichier CSV\n",
    "df.to_csv(\"descriptor_GetAvailableProperties.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "907fe08a-389f-4b7b-972f-42fd64babb8f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1245, 48)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_2.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "1c1c7593-ecdd-4e40-8935-a5ed20e200fd",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SMILES                             0\n",
       "yield                              0\n",
       "Temperature                        0\n",
       "Ind                                0\n",
       "SMILES_2                           0\n",
       "mols                               0\n",
       "amw                                0\n",
       "lipinskiHBA                        0\n",
       "lipinskiHBD                        0\n",
       "NumRotatableBonds                  0\n",
       "NumHBD                             0\n",
       "NumHBA                             0\n",
       "NumHeavyAtoms                      0\n",
       "NumAtoms                           0\n",
       "NumHeteroatoms                     0\n",
       "NumAmideBonds                      0\n",
       "FractionCSP3                       0\n",
       "NumRings                           0\n",
       "NumAromaticRings                   0\n",
       "NumAliphaticRings                  0\n",
       "NumSaturatedRings                  0\n",
       "NumHeterocycles                    0\n",
       "NumAromaticHeterocycles            0\n",
       "NumSaturatedHeterocycles           0\n",
       "NumAliphaticHeterocycles           0\n",
       "NumSpiroAtoms                      0\n",
       "NumBridgeheadAtoms                 0\n",
       "NumAtomStereoCenters               0\n",
       "NumUnspecifiedAtomStereoCenters    0\n",
       "labuteASA                          0\n",
       "tpsa                               0\n",
       "CrippenClogP                       0\n",
       "CrippenMR                          0\n",
       "chi0v                              0\n",
       "chi1v                              0\n",
       "chi2v                              0\n",
       "chi3v                              0\n",
       "chi4v                              0\n",
       "chi0n                              0\n",
       "chi1n                              0\n",
       "chi2n                              0\n",
       "chi3n                              0\n",
       "chi4n                              0\n",
       "hallKierAlpha                      0\n",
       "kappa1                             0\n",
       "kappa2                             0\n",
       "kappa3                             0\n",
       "Phi                                0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_2.isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "c86c98db-c136-4b5c-ba49-92fa3a13ae6c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['SMILES', 'yield', 'Temperature', 'Ind', 'SMILES_2', 'mols', 'amw',\n",
       "       'lipinskiHBA', 'lipinskiHBD', 'NumRotatableBonds', 'NumHBD', 'NumHBA',\n",
       "       'NumHeavyAtoms', 'NumAtoms', 'NumHeteroatoms', 'NumAmideBonds',\n",
       "       'FractionCSP3', 'NumRings', 'NumAromaticRings', 'NumAliphaticRings',\n",
       "       'NumSaturatedRings', 'NumHeterocycles', 'NumAromaticHeterocycles',\n",
       "       'NumSaturatedHeterocycles', 'NumAliphaticHeterocycles', 'NumSpiroAtoms',\n",
       "       'NumBridgeheadAtoms', 'NumAtomStereoCenters',\n",
       "       'NumUnspecifiedAtomStereoCenters', 'labuteASA', 'tpsa', 'CrippenClogP',\n",
       "       'CrippenMR', 'chi0v', 'chi1v', 'chi2v', 'chi3v', 'chi4v', 'chi0n',\n",
       "       'chi1n', 'chi2n', 'chi3n', 'chi4n', 'hallKierAlpha', 'kappa1', 'kappa2',\n",
       "       'kappa3', 'Phi'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_2.columns"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "52ff2366-6af5-4d08-b804-ebf17598d21c",
   "metadata": {},
   "source": [
    "### Filtrage des features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "d479abe4-96a6-47c9-bd3a-54a2f1f75150",
   "metadata": {},
   "outputs": [],
   "source": [
    "def filter_descriptors_low(data, threshold):\n",
    "    # Drop colonne inutile\n",
    "    data = data.drop([\"SMILES\", \"Ind\", \"SMILES_2\", \"mols\"], axis=1)\n",
    "    # Copie des descripteurs\n",
    "    descriptors = data.copy()\n",
    "    \n",
    "    # Calcul de la matrice de corrélation et mise à zéro de la diagonale\n",
    "    corr = descriptors.corr()\n",
    "    for index in corr.index:\n",
    "        corr.loc[index, index] = 0\n",
    "    \n",
    "    # Préparation d'une liste pour collecter les noms des descripteurs non fortement corrélés\n",
    "    descriptors_not_correlated = []\n",
    "    \n",
    "    # Vérifier chaque colonne individuellement\n",
    "    for col in corr.columns:\n",
    "        if all((corr[col][corr[col].notna()] <= threshold)):\n",
    "            descriptors_not_correlated.append(col)\n",
    "    \n",
    "    print('Number of descriptors:', len(descriptors_not_correlated))\n",
    "    \n",
    "    # Retourner les données filtrées\n",
    "    return data[descriptors_not_correlated]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b3e8869b-a93c-4741-8669-04c64df29b79",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "cafb74ea-3eb3-471e-9c73-756f6dc8d930",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of descriptors: 15\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(1245, 15)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_3_desc = filter_descriptors_low(df_2, 0.9)\n",
    "df_3_desc.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "607bd0f5-a751-4b90-9050-7e0012c1d97c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['yield', 'Temperature', 'lipinskiHBD', 'NumHBD', 'NumHeteroatoms',\n",
       "       'NumAmideBonds', 'FractionCSP3', 'NumAliphaticRings',\n",
       "       'NumSaturatedRings', 'NumSaturatedHeterocycles',\n",
       "       'NumAliphaticHeterocycles', 'NumSpiroAtoms', 'NumAtomStereoCenters',\n",
       "       'NumUnspecifiedAtomStereoCenters', 'hallKierAlpha'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_3_desc.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "32065c85-1601-487d-86b8-282a897e2945",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "def filter_descriptors_hard(data, threshold):\n",
    "    # Drop colonne inutile\n",
    "    data = data.drop([\"SMILES\", \"Ind\", \"SMILES_2\", \"mols\"], axis=1)\n",
    "    # Copie des descripteurs\n",
    "    descriptors = data.copy()\n",
    "    \n",
    "    # Calcul de la matrice de corrélation et mise à zéro de la diagonale\n",
    "    corr = descriptors.corr()\n",
    "    for index in corr.index:\n",
    "        corr.loc[index, index] = 0\n",
    "    \n",
    "    # Identifie les groupes de descripteurs fortement corrélés\n",
    "    corr_mask = corr.abs() > threshold\n",
    "    groups = []\n",
    "    for col in corr.columns:\n",
    "        correlated_with = corr.index[corr_mask[col]].tolist()\n",
    "        if correlated_with and col not in sum(groups, []):\n",
    "            groups.append(correlated_with)\n",
    "    \n",
    "    # Choix d'un descripteur par groupe\n",
    "    selected_descriptors = []\n",
    "    for group in groups:\n",
    "        # Choix basé sur le minimum de corrélations maximales avec les autres\n",
    "        min_correlation_sum = np.inf\n",
    "        selected = None\n",
    "        for item in group:\n",
    "            # Assurez-vous de ne considérer que les valeurs non-NaN pour la somme des corrélations\n",
    "            corr_sum = corr[item][corr_mask[item] & corr[item].notna()].sum()\n",
    "            if corr_sum < min_correlation_sum:\n",
    "                min_correlation_sum = corr_sum\n",
    "                selected = item\n",
    "        selected_descriptors.append(selected)\n",
    "    \n",
    "    print('Descripteurs sélectionnés:', selected_descriptors)\n",
    "        \n",
    "    # Retourner les données filtrées\n",
    "    return data[selected_descriptors], data['yield']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "d6a59e5e-0650-4478-9fe3-0008bdb304d1",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Descripteurs sélectionnés: ['chi4v', 'NumHBA', 'kappa3', 'hallKierAlpha', 'NumAromaticHeterocycles', 'kappa3']\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((1245, 6), (1245,))"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_5, df_5_yield = filter_descriptors_hard(df_2, 0.9)\n",
    "df_5.shape, df_5_yield.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "93bb56ab-8a47-44be-94ab-4601df486eec",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "611ae8ad-6ba9-4276-9d0c-966da9be108a",
   "metadata": {},
   "source": [
    "## Model "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "67049a73-fa43-4349-93c8-53614cd508a3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "from sklearn import metrics\n",
    "from sklearn.model_selection import cross_val_score, KFold, train_test_split\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from sklearn.model_selection import train_test_split, RandomizedSearchCV\n",
    "from sklearn.metrics import mean_squared_error, r2_score, mean_absolute_error\n",
    "from statistics import mean, stdev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c9aa6bc4-3cd9-409b-864d-199b9d26f58a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "d9f6e99c-1676-4ab0-af6c-5a7c9d3ee008",
   "metadata": {},
   "source": [
    "### Model_random_forest_regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "28d21bb9-bf23-4d05-a835-b2648840b00e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def random_forest_regression(X, y):\n",
    "    rmse_scores = []\n",
    "    r2_scores = []\n",
    "    for i in range(10):\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        model = RandomForestRegressor(n_estimators=120, max_depth=25)\n",
    "        model.fit(X_train, y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "        rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "        rmse_scores.append(rmse)\n",
    "        r2 = r2_score(y_test, y_pred)\n",
    "        r2_scores.append(r2)\n",
    "        #print(f\"Iteration {i}: RMSE = {rmse:.3f}, R^2  = {r2:.3f}\")\n",
    "    # Retourner la moyenne et l'écart-type des scores\n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)\n",
    "    \n",
    "    print(\"\\nAverage performance of test set at 10 split\")\n",
    "    print(\"Metric\\tavg\\tstdev\")\n",
    "    print(\"R^2\\t%.2f\\t%.2f\" % (mean(r2_scores), stdev(r2_scores)))\n",
    "    print(\"RMSE\\t%.2f\\t%.2f\" % (mean(rmse_scores), stdev(rmse_scores)))"
   ]
  },
  {
   "cell_type": "raw",
   "id": "8ece8524-fbca-43bf-bc4e-370a0d78c74e",
   "metadata": {},
   "source": [
    "def statistics_for_the_training_set_with_5CV_RFR(X, y):\n",
    "    r2_scores = []\n",
    "    rmse_scores = []\n",
    "    for i in range(10):\n",
    "        X_tr, X_t, y_tr, y_t = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        X_tr.reset_index(drop=True , inplace=True)\n",
    "        y_tr.reset_index(drop=True , inplace=True)\n",
    "        cv = KFold(n_splits=5, shuffle=True, random_state=1)\n",
    "        y_pred, y_true = [], []\n",
    "        for train_index, test_index in cv.split(X_tr):\n",
    "            X_train = X_tr.loc[train_index].values\n",
    "            X_test = X_tr.loc[test_index].values\n",
    "            y_train = y_tr.loc[train_index].values\n",
    "            y_test = y_tr.loc[test_index].values\n",
    "    \n",
    "            model = RandomForestRegressor(n_estimators=120, max_depth=25)\n",
    "            model.fit(X_train, y_train)\n",
    "            y_pred.extend(model.predict(X_test))\n",
    "            y_true.extend(y_test)\n",
    "\n",
    "            # Calculer les métriques\n",
    "            rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "            r2 = r2_score(y_test, y_pred)\n",
    "\n",
    "        r2_scores.append(metrics.r2_score(y_true, y_pred))\n",
    "        rmse_scores.append(metrics.mean_squared_error(y_true, y_pred, squared=False))\n",
    "        \n",
    "    # Retourner la moyenne et l'écart-type des scores\n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)   \n",
    "\n",
    "    print(\"\\nAverage performance of 5CV\")\n",
    "    print(\"Metric\\tavg\\tstdev\")\n",
    "    print(\"R^2\\t%.2f\\t%.2f\" % (mean(r2_scores), stdev(r2_scores)))\n",
    "    print(\"RMSE\\t%.2f\\t%.2f\" % (mean(rmse_scores), stdev(rmse_scores)))   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "82f188ea-257e-4340-8525-ccd2e4b1cec0",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from statistics import mean, stdev\n",
    "\n",
    "def statistics_for_the_training_set_with_5CV_RFR(X, y):\n",
    "    r2_scores = []\n",
    "    rmse_scores = []\n",
    "    \n",
    "    for i in range(10):\n",
    "        X_tr, X_t, y_tr, y_t = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        X_tr.reset_index(drop=True, inplace=True)\n",
    "        y_tr.reset_index(drop=True, inplace=True)\n",
    "        \n",
    "        cv = KFold(n_splits=5, shuffle=True, random_state=1)\n",
    "        fold_r2_scores = []\n",
    "        fold_rmse_scores = []\n",
    "        \n",
    "        for train_index, test_index in cv.split(X_tr):\n",
    "            X_train, X_test = X_tr.loc[train_index].values, X_tr.loc[test_index].values\n",
    "            y_train, y_test = y_tr.loc[train_index].values, y_tr.loc[test_index].values\n",
    "            \n",
    "            model = RandomForestRegressor(n_estimators=120, max_depth=25)\n",
    "            model.fit(X_train, y_train)\n",
    "            y_pred = model.predict(X_test)\n",
    "            \n",
    "            fold_rmse_scores.append(np.sqrt(mean_squared_error(y_test, y_pred)))\n",
    "            fold_r2_scores.append(r2_score(y_test, y_pred))\n",
    "        \n",
    "        # Moyennes et écarts-types des scores pour chaque pli\n",
    "        rmse_scores.append(mean(fold_rmse_scores))\n",
    "        r2_scores.append(mean(fold_r2_scores))\n",
    "    \n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)\n"
   ]
  },
  {
   "cell_type": "raw",
   "id": "8ad6facb-1487-43b1-a562-cd8ad9a76065",
   "metadata": {},
   "source": [
    "# The \"reaction\" strategy\n",
    "print(\"Random Forest Regression:\")\n",
    "random_forest_regression(X, y)\n",
    "\n",
    "print(\"\\nStatistics for the training set with 5CV RFR:\")\n",
    "statistics_for_the_training_set_with_5CV_RFR(X, y)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6637a577-961c-48b1-8bfe-c02b48895a38",
   "metadata": {},
   "source": [
    "### Model CatBoost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "15c040e5-81a8-4378-99d4-bb6a276b4100",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Importer CatBoost après l'installation\n",
    "import catboost"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cdaac37b-1c7c-4e8d-9cae-6a50498edd36",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "306d3f89-e133-482a-b61b-cbbbd4d85d7a",
   "metadata": {},
   "outputs": [],
   "source": [
    "def catboost_regression(X, y):\n",
    "    rmse_scores = []\n",
    "    r2_scores = []\n",
    "    for i in range(10):\n",
    "        X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        model = catboost.CatBoostRegressor(learning_rate=0.02, l2_leaf_reg=9.5, depth=7, silent=True)\n",
    "        model.fit(X_train, y_train)\n",
    "        y_pred = model.predict(X_test)\n",
    "        rmse = np.sqrt(mean_squared_error(y_test, y_pred))\n",
    "        rmse_scores.append(rmse)\n",
    "        r2 = r2_score(y_test, y_pred)\n",
    "        r2_scores.append(r2)\n",
    "        #print(f\"Iteration {i}: RMSE = {rmse:.3f}, R2 = {r2:.3f}\")\n",
    "    \n",
    "    # Retourner la moyenne et l'écart-type des scores\n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)\n",
    "    \n",
    "    print(\"\\nAverage performance of test set at 10 split\")\n",
    "    print(\"Metric\\tavg\\tstdev\")\n",
    "    print(\"R^2\\t%.2f\\t%.2f\" % (mean(r2_scores), stdev(r2_scores)))\n",
    "    print(\"RMSE\\t%.2f\\t%.2f\" % (mean(rmse_scores), stdev(rmse_scores)))"
   ]
  },
  {
   "cell_type": "raw",
   "id": "1583c50c-3df0-415b-a23d-a5bfa691285a",
   "metadata": {},
   "source": [
    "def statistics_for_the_training_set_with_5CV_catBoost(X, y):\n",
    "    r2_scores = []\n",
    "    rmse_scores = []\n",
    "    for i in range(10):\n",
    "        X_tr, X_t, y_tr, y_t = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        X_tr.reset_index(drop=True , inplace=True)\n",
    "        y_tr.reset_index(drop=True , inplace=True)\n",
    "        cv = KFold(n_splits=5, shuffle=True, random_state=1)\n",
    "        y_pred, y_true = [], []\n",
    "        for train_index, test_index in cv.split(X_tr):\n",
    "            X_train = X_tr.loc[train_index].values\n",
    "            X_test = X_tr.loc[test_index].values\n",
    "            y_train = y_tr.loc[train_index].values\n",
    "            y_test = y_tr.loc[test_index].values\n",
    "    \n",
    "            model = catboost.CatBoostRegressor(learning_rate=0.02, l2_leaf_reg=9.5, depth=7, silent=True)\n",
    "            model.fit(X_train, y_train)\n",
    "            y_pred.extend(model.predict(X_test))\n",
    "            y_true.extend(y_test)\n",
    "\n",
    "        r2_scores.append(metrics.r2_score(y_true, y_pred))\n",
    "        rmse_scores.append(metrics.mean_squared_error(y_true, y_pred, squared=False))\n",
    "        \n",
    "    # Retourner la moyenne et l'écart-type des scores\n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)  \n",
    "    print(\"\\nAverage performance of 5CV\")\n",
    "    print(\"Metric\\tavg\\tstdev\")\n",
    "    print(\"R^2\\t%.2f\\t%.2f\" % (mean(r2_scores), stdev(r2_scores)))\n",
    "    print(\"RMSE\\t%.2f\\t%.2f\" % (mean(rmse_scores), stdev(rmse_scores)))        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "id": "8025095d-f9d6-4ce1-bb25-72e6487785c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold\n",
    "from sklearn.metrics import mean_squared_error, r2_score\n",
    "from sklearn.ensemble import RandomForestRegressor\n",
    "from statistics import mean, stdev\n",
    "\n",
    "def statistics_for_the_training_set_with_5CV_catBoost(X, y):\n",
    "    r2_scores = []\n",
    "    rmse_scores = []\n",
    "    \n",
    "    for i in range(10):\n",
    "        X_tr, X_t, y_tr, y_t = train_test_split(X, y, test_size=0.2, random_state=i)\n",
    "        X_tr.reset_index(drop=True, inplace=True)\n",
    "        y_tr.reset_index(drop=True, inplace=True)\n",
    "        \n",
    "        cv = KFold(n_splits=5, shuffle=True, random_state=1)\n",
    "        fold_r2_scores = []\n",
    "        fold_rmse_scores = []\n",
    "        \n",
    "        for train_index, test_index in cv.split(X_tr):\n",
    "            X_train, X_test = X_tr.loc[train_index].values, X_tr.loc[test_index].values\n",
    "            y_train, y_test = y_tr.loc[train_index].values, y_tr.loc[test_index].values\n",
    "            \n",
    "            model = catboost.CatBoostRegressor(learning_rate=0.02, l2_leaf_reg=9.5, depth=7, silent=True)\n",
    "            model.fit(X_train, y_train)\n",
    "            y_pred = model.predict(X_test)\n",
    "            \n",
    "            fold_rmse_scores.append(np.sqrt(mean_squared_error(y_test, y_pred)))\n",
    "            fold_r2_scores.append(r2_score(y_test, y_pred))\n",
    "        \n",
    "        # Moyennes et écarts-types des scores pour chaque pli\n",
    "        rmse_scores.append(mean(fold_rmse_scores))\n",
    "        r2_scores.append(mean(fold_r2_scores))\n",
    "    \n",
    "    return mean(rmse_scores), stdev(rmse_scores), mean(r2_scores), stdev(r2_scores)\n"
   ]
  },
  {
   "cell_type": "raw",
   "id": "ded1b84e-c67b-4ba7-b38f-759dc2d96b50",
   "metadata": {},
   "source": [
    "# The \"reaction\" strategy\n",
    "print(\"\\nCatBoost Regression:\")\n",
    "catboost_regression(X, y)\n",
    "    \n",
    "print(\"\\nStatistics for the training set with 5CV catBoost:\")\n",
    "statistics_for_the_training_set_with_5CV_catBoost(X, y)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9e37444a-afbd-47e2-99ab-1bb7fed50647",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "639f8f11-28c7-4cc6-a201-b5a394a44e7d",
   "metadata": {},
   "source": [
    "## Result Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "832c7f6a-f71f-43c0-b50e-254ba61bc515",
   "metadata": {},
   "source": [
    "### Result Model df_3_LOW_filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "id": "2c38f71f-2696-45a3-8e79-b66cfdf1b05f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1245, 15)"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_3_desc.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "8e2b173f-245b-4689-9fa3-d576bb8d0900",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_3_desc.isnull().sum().max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "e9b4f790-06cf-461e-a564-a35c7933911b",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>yield</th>\n",
       "      <th>Temperature</th>\n",
       "      <th>lipinskiHBD</th>\n",
       "      <th>NumHBD</th>\n",
       "      <th>NumHeteroatoms</th>\n",
       "      <th>NumAmideBonds</th>\n",
       "      <th>FractionCSP3</th>\n",
       "      <th>NumAliphaticRings</th>\n",
       "      <th>NumSaturatedRings</th>\n",
       "      <th>NumSaturatedHeterocycles</th>\n",
       "      <th>NumAliphaticHeterocycles</th>\n",
       "      <th>NumSpiroAtoms</th>\n",
       "      <th>NumAtomStereoCenters</th>\n",
       "      <th>NumUnspecifiedAtomStereoCenters</th>\n",
       "      <th>hallKierAlpha</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.05</td>\n",
       "      <td>20</td>\n",
       "      <td>12.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.360544</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-17.300000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.00</td>\n",
       "      <td>40</td>\n",
       "      <td>10.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>37.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.130435</td>\n",
       "      <td>5.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-16.685065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2.00</td>\n",
       "      <td>20</td>\n",
       "      <td>10.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.176471</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-10.535065</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2.10</td>\n",
       "      <td>20</td>\n",
       "      <td>12.0</td>\n",
       "      <td>8.0</td>\n",
       "      <td>27.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.377483</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>6.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-17.300000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2.50</td>\n",
       "      <td>25</td>\n",
       "      <td>11.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.062992</td>\n",
       "      <td>3.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-19.560000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   yield  Temperature  lipinskiHBD  NumHBD  NumHeteroatoms  NumAmideBonds  \\\n",
       "0   0.05           20         12.0     8.0            27.0            0.0   \n",
       "1   1.00           40         10.0     6.0            37.0            0.0   \n",
       "2   2.00           20         10.0     6.0            30.0            0.0   \n",
       "3   2.10           20         12.0     8.0            27.0            0.0   \n",
       "4   2.50           25         11.0     7.0            45.0            0.0   \n",
       "\n",
       "   FractionCSP3  NumAliphaticRings  NumSaturatedRings  \\\n",
       "0      0.360544                6.0                0.0   \n",
       "1      0.130435                5.0                0.0   \n",
       "2      0.176471                3.0                0.0   \n",
       "3      0.377483                6.0                0.0   \n",
       "4      0.062992                3.0                0.0   \n",
       "\n",
       "   NumSaturatedHeterocycles  NumAliphaticHeterocycles  NumSpiroAtoms  \\\n",
       "0                       0.0                       6.0            0.0   \n",
       "1                       0.0                       4.0            0.0   \n",
       "2                       0.0                       2.0            0.0   \n",
       "3                       0.0                       6.0            0.0   \n",
       "4                       0.0                       2.0            0.0   \n",
       "\n",
       "   NumAtomStereoCenters  NumUnspecifiedAtomStereoCenters  hallKierAlpha  \n",
       "0                   0.0                              0.0     -17.300000  \n",
       "1                   0.0                              0.0     -16.685065  \n",
       "2                   0.0                              0.0     -10.535065  \n",
       "3                   0.0                              0.0     -17.300000  \n",
       "4                   0.0                              0.0     -19.560000  "
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_3_desc.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "495d2fc5-7f30-433e-a2a1-3c6de908bfab",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Présence de valeurs infinies dans chaque colonne :\n",
      "yield                              False\n",
      "Temperature                        False\n",
      "lipinskiHBD                        False\n",
      "NumHBD                             False\n",
      "NumHeteroatoms                     False\n",
      "NumAmideBonds                      False\n",
      "FractionCSP3                       False\n",
      "NumAliphaticRings                  False\n",
      "NumSaturatedRings                  False\n",
      "NumSaturatedHeterocycles           False\n",
      "NumAliphaticHeterocycles           False\n",
      "NumSpiroAtoms                      False\n",
      "NumAtomStereoCenters               False\n",
      "NumUnspecifiedAtomStereoCenters    False\n",
      "hallKierAlpha                      False\n",
      "dtype: bool\n"
     ]
    }
   ],
   "source": [
    "# Vérifier s'il y a des valeurs infinies dans le DataFrame\n",
    "has_infinities = np.isinf(df_3_desc).any()\n",
    "\n",
    "print(\"Présence de valeurs infinies dans chaque colonne :\")\n",
    "print(has_infinities)\n"
   ]
  },
  {
   "cell_type": "raw",
   "id": "b9ee92c8-0d1f-43f0-b7cd-5b0ae958afd0",
   "metadata": {},
   "source": [
    "# Supprimer colonne IPC, car dans X, nous voyons une valeur infinie\n",
    "df_4_desc = df_3_desc.drop([\"Ipc\"], axis=1)"
   ]
  },
  {
   "cell_type": "raw",
   "id": "ae3762ed-69d2-4ec9-8b44-4de4cdf78edd",
   "metadata": {},
   "source": [
    "df_4_desc.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "id": "fd4de7fd-8210-4f63-b7d9-0664df39c13d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1245, 14) (1245,)\n"
     ]
    }
   ],
   "source": [
    "X=df_3_desc.drop([\"yield\"], axis=1)\n",
    "y=df_3_desc['yield']\n",
    "\n",
    "print(X.shape, y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "c080d3b2-2b79-4214-ba76-ef9929a5ad39",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Maximum values in each column:\n",
      " Temperature                        360.000000\n",
      "lipinskiHBD                         20.000000\n",
      "NumHBD                              14.000000\n",
      "NumHeteroatoms                      73.000000\n",
      "NumAmideBonds                        4.000000\n",
      "FractionCSP3                         0.765766\n",
      "NumAliphaticRings                   23.000000\n",
      "NumSaturatedRings                    8.000000\n",
      "NumSaturatedHeterocycles             6.000000\n",
      "NumAliphaticHeterocycles             6.000000\n",
      "NumSpiroAtoms                        1.000000\n",
      "NumAtomStereoCenters                17.000000\n",
      "NumUnspecifiedAtomStereoCenters     17.000000\n",
      "hallKierAlpha                        0.852597\n",
      "dtype: float64\n",
      "Minimum values in each column:\n",
      " Temperature                       -78.000000\n",
      "lipinskiHBD                         0.000000\n",
      "NumHBD                              0.000000\n",
      "NumHeteroatoms                      6.000000\n",
      "NumAmideBonds                       0.000000\n",
      "FractionCSP3                        0.037383\n",
      "NumAliphaticRings                   0.000000\n",
      "NumSaturatedRings                   0.000000\n",
      "NumSaturatedHeterocycles            0.000000\n",
      "NumAliphaticHeterocycles            0.000000\n",
      "NumSpiroAtoms                       0.000000\n",
      "NumAtomStereoCenters                0.000000\n",
      "NumUnspecifiedAtomStereoCenters     0.000000\n",
      "hallKierAlpha                     -22.005065\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(\"Maximum values in each column:\\n\", X.max())\n",
    "print(\"Minimum values in each column:\\n\", X.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "2030e9ac-c1bd-40bd-8b16-afe9e65bc9c4",
   "metadata": {},
   "outputs": [],
   "source": [
    "import time\n",
    "import pandas as pd\n",
    "from tqdm import tqdm\n",
    "\n",
    "def run_and_compare_models(X, y):\n",
    "    models = {\n",
    "        \"Random_Forest\": random_forest_regression,\n",
    "        \"Random_Forest_Kfold\": statistics_for_the_training_set_with_5CV_RFR,\n",
    "        \"CatBoost\": catboost_regression,\n",
    "        \"CatBoostKflod\": statistics_for_the_training_set_with_5CV_catBoost\n",
    "    }\n",
    "    \n",
    "    results = []\n",
    "    \n",
    "    for model_name, model_func in tqdm(models.items(), desc=\"Running models\"):\n",
    "        start_time = time.time()\n",
    "        rmse_avg, rmse_std, r2_avg, r2_std = model_func(X, y)\n",
    "        end_time = time.time()  # Arrêter le chronomètre\n",
    "        elapsed_time = end_time - start_time\n",
    "        results.append({\n",
    "            \"Model\": model_name,\n",
    "            \"RMSE Avg\": rmse_avg,\n",
    "            \"RMSE Std\": rmse_std,\n",
    "            \"R^2 Avg\": r2_avg,\n",
    "            \"R^2 Std\": r2_std,\n",
    "            \"Time (s)\": elapsed_time  # Ajouter le temps écoulé à chaque entrée\n",
    "        })\n",
    "        \n",
    "    results_df = pd.DataFrame(results)\n",
    "    return results_df\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "4f8255ef-7646-4252-b473-4d5ba9433b85",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Running models: 100%|██████████| 4/4 [01:48<00:00, 27.13s/it]\n"
     ]
    }
   ],
   "source": [
    "results_df_ = run_and_compare_models(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "7ec53389-9c00-4dc4-a395-b40701a4bd0b",
   "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>Model</th>\n",
       "      <th>RMSE Avg</th>\n",
       "      <th>RMSE Std</th>\n",
       "      <th>R^2 Avg</th>\n",
       "      <th>R^2 Std</th>\n",
       "      <th>Time (s)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Random_Forest</td>\n",
       "      <td>14.138347</td>\n",
       "      <td>0.750968</td>\n",
       "      <td>0.552160</td>\n",
       "      <td>0.048000</td>\n",
       "      <td>3.564363</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Random_Forest_Kfold</td>\n",
       "      <td>14.357041</td>\n",
       "      <td>0.268192</td>\n",
       "      <td>0.543474</td>\n",
       "      <td>0.019621</td>\n",
       "      <td>14.892357</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CatBoost</td>\n",
       "      <td>13.871584</td>\n",
       "      <td>0.662638</td>\n",
       "      <td>0.568857</td>\n",
       "      <td>0.044257</td>\n",
       "      <td>14.334295</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CatBoostKflod</td>\n",
       "      <td>14.126811</td>\n",
       "      <td>0.182020</td>\n",
       "      <td>0.558135</td>\n",
       "      <td>0.017690</td>\n",
       "      <td>75.742743</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Model   RMSE Avg  RMSE Std   R^2 Avg   R^2 Std   Time (s)\n",
       "0        Random_Forest  14.138347  0.750968  0.552160  0.048000   3.564363\n",
       "1  Random_Forest_Kfold  14.357041  0.268192  0.543474  0.019621  14.892357\n",
       "2             CatBoost  13.871584  0.662638  0.568857  0.044257  14.334295\n",
       "3        CatBoostKflod  14.126811  0.182020  0.558135  0.017690  75.742743"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "cf162c9c-a94a-48cb-a8c2-d32d630ca373",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Sauvegarde du DataFrame dans un fichier CSV\n",
    "results_df_.to_csv(\"Resultat_Df_filtrage_Low_descriptor_GetAvailableProperties.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aeafb12d-3adf-472d-9861-333471ea0151",
   "metadata": {
    "jp-MarkdownHeadingCollapsed": true
   },
   "source": [
    "### Result Model df_5_HARD_filter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "87fed1c3-daf1-4a6b-9f1d-352633cdf321",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((1245, 6), (1245,))"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_5.shape, df_5_yield.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "ecdba950-01a0-4ab9-9de1-5e3fd5c64e69",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1245, 6) (1245,)\n"
     ]
    }
   ],
   "source": [
    "X=df_5\n",
    "y= df_5_yield\n",
    "\n",
    "print(X.shape, y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "b8c647e0-c712-4ea3-bc93-d905ce42c65f",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Maximum values in each column:\n",
      " chi4v                      52.424356\n",
      "NumHBA                     41.000000\n",
      "kappa3                     87.419036\n",
      "hallKierAlpha               0.852597\n",
      "NumAromaticHeterocycles    15.000000\n",
      "kappa3                     87.419036\n",
      "dtype: float64\n",
      "Minimum values in each column:\n",
      " chi4v                       2.712759\n",
      "NumHBA                      1.000000\n",
      "kappa3                      9.195023\n",
      "hallKierAlpha             -22.005065\n",
      "NumAromaticHeterocycles     0.000000\n",
      "kappa3                      9.195023\n",
      "dtype: float64\n"
     ]
    }
   ],
   "source": [
    "print(\"Maximum values in each column:\\n\", X.max())\n",
    "print(\"Minimum values in each column:\\n\", X.min())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "c3bea452-4039-44b3-ac0e-5843fce1dbac",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = X.loc[:, ~X.columns.duplicated()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "ae485ba2-d7de-45f4-bec7-c6e86504db42",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['chi4v', 'NumHBA', 'kappa3', 'hallKierAlpha',\n",
       "       'NumAromaticHeterocycles'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "3c23593f-d7ca-4b4f-a67f-33c5f11b3729",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1245, 5) (1245,)\n"
     ]
    }
   ],
   "source": [
    "print(X.shape, y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "ae56f19f-11ff-4983-8001-b4de6ed3f7d1",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Running models: 100%|██████████| 4/4 [01:43<00:00, 25.78s/it]\n"
     ]
    }
   ],
   "source": [
    "results_df_1 = run_and_compare_models(X, y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "b81eb958-ad7b-4a5c-adf4-864125a45487",
   "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>Model</th>\n",
       "      <th>RMSE Avg</th>\n",
       "      <th>RMSE Std</th>\n",
       "      <th>R^2 Avg</th>\n",
       "      <th>R^2 Std</th>\n",
       "      <th>Time (s)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Random_Forest</td>\n",
       "      <td>16.424956</td>\n",
       "      <td>0.843989</td>\n",
       "      <td>0.396642</td>\n",
       "      <td>0.052442</td>\n",
       "      <td>3.537553</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Random_Forest_Kfold</td>\n",
       "      <td>16.441462</td>\n",
       "      <td>0.346823</td>\n",
       "      <td>0.401929</td>\n",
       "      <td>0.026976</td>\n",
       "      <td>14.694441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CatBoost</td>\n",
       "      <td>15.797854</td>\n",
       "      <td>0.706779</td>\n",
       "      <td>0.441718</td>\n",
       "      <td>0.045898</td>\n",
       "      <td>13.901132</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CatBoostKflod</td>\n",
       "      <td>15.859477</td>\n",
       "      <td>0.214479</td>\n",
       "      <td>0.443892</td>\n",
       "      <td>0.016753</td>\n",
       "      <td>70.977397</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 Model   RMSE Avg  RMSE Std   R^2 Avg   R^2 Std   Time (s)\n",
       "0        Random_Forest  16.424956  0.843989  0.396642  0.052442   3.537553\n",
       "1  Random_Forest_Kfold  16.441462  0.346823  0.401929  0.026976  14.694441\n",
       "2             CatBoost  15.797854  0.706779  0.441718  0.045898  13.901132\n",
       "3        CatBoostKflod  15.859477  0.214479  0.443892  0.016753  70.977397"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results_df_1 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "1ad28223-194f-4ea7-b655-b86a9ecc7dd8",
   "metadata": {},
   "outputs": [],
   "source": [
    "results_df_1.to_csv('Resultat_Df_filtrage_HARD_descriptor_GetAvailableProperties.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b7296712-502d-4e88-a6ef-a392858c8962",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "100jML",
   "language": "python",
   "name": "100jml"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.12.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}