{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "X5AMaKsOmkjy"
},
"source": [
"## 1) Problem statement.\n",
"Phase prediction of HEAs for SMOTETomek links augmented data"
]
},
{
"cell_type": "code",
"source": [
"pip install miceforest ###for MICE imputer"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "0BnBo3LO4u8o",
"outputId": "d7c82579-4650-47a9-c5c8-02cb76d85426"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
"Collecting miceforest\n",
" Downloading miceforest-5.6.3-py3-none-any.whl (57 kB)\n",
"\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m58.0/58.0 KB\u001b[0m \u001b[31m3.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
"\u001b[?25hRequirement already satisfied: numpy in /usr/local/lib/python3.8/dist-packages (from miceforest) (1.21.6)\n",
"Requirement already satisfied: dill in /usr/local/lib/python3.8/dist-packages (from miceforest) (0.3.6)\n",
"Collecting lightgbm>=3.3.1\n",
" Downloading lightgbm-3.3.5-py3-none-manylinux1_x86_64.whl (2.0 MB)\n",
"\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.0/2.0 MB\u001b[0m \u001b[31m34.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
"\u001b[?25hCollecting blosc\n",
" Downloading blosc-1.11.1-cp38-cp38-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (2.5 MB)\n",
"\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.5/2.5 MB\u001b[0m \u001b[31m59.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
"\u001b[?25hRequirement already satisfied: wheel in /usr/local/lib/python3.8/dist-packages (from lightgbm>=3.3.1->miceforest) (0.38.4)\n",
"Requirement already satisfied: scipy in /usr/local/lib/python3.8/dist-packages (from lightgbm>=3.3.1->miceforest) (1.7.3)\n",
"Requirement already satisfied: scikit-learn!=0.22.0 in /usr/local/lib/python3.8/dist-packages (from lightgbm>=3.3.1->miceforest) (1.0.2)\n",
"Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.8/dist-packages (from scikit-learn!=0.22.0->lightgbm>=3.3.1->miceforest) (3.1.0)\n",
"Requirement already satisfied: joblib>=0.11 in /usr/local/lib/python3.8/dist-packages (from scikit-learn!=0.22.0->lightgbm>=3.3.1->miceforest) (1.2.0)\n",
"Installing collected packages: blosc, lightgbm, miceforest\n",
" Attempting uninstall: lightgbm\n",
" Found existing installation: lightgbm 2.2.3\n",
" Uninstalling lightgbm-2.2.3:\n",
" Successfully uninstalled lightgbm-2.2.3\n",
"Successfully installed blosc-1.11.1 lightgbm-3.3.5 miceforest-5.6.3\n"
]
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "OYaew9gYmkj1"
},
"source": [
"## 2) Import required libraries"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "SG7iLs0Zmkj1"
},
"outputs": [],
"source": [
"import pandas as pd\n",
"import seaborn as sns\n",
"import numpy as np\n",
"from statistics import mean\n",
"import matplotlib.pyplot as plt\n",
"import warnings\n",
"from sklearn.preprocessing import PowerTransformer\n",
"import numpy as np\n",
"from sklearn.preprocessing import LabelEncoder\n",
"from sklearn.pipeline import Pipeline\n",
"from sklearn.utils import resample\n",
"\n",
"from sklearn.linear_model import LogisticRegression\n",
"from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier, GradientBoostingClassifier\n",
"from sklearn.neighbors import KNeighborsClassifier\n",
"from sklearn.tree import DecisionTreeClassifier\n",
"from sklearn.svm import SVC\n",
"from sklearn.metrics import accuracy_score, classification_report,ConfusionMatrixDisplay, \\\n",
" precision_score, recall_score, f1_score, roc_auc_score,roc_curve,confusion_matrix\n",
"\n",
"\n",
"from sklearn import metrics \n",
"from sklearn.model_selection import train_test_split, RepeatedStratifiedKFold, cross_val_score\n",
"from sklearn.preprocessing import OneHotEncoder, MinMaxScaler\n",
"from sklearn.compose import ColumnTransformer\n",
"from sklearn.impute import SimpleImputer, KNNImputer\n",
"from xgboost import XGBClassifier\n",
"from sklearn.preprocessing import StandardScaler, MinMaxScaler,RobustScaler\n",
"from sklearn.compose import ColumnTransformer\n",
"\n",
"warnings.filterwarnings(\"ignore\")\n",
"%matplotlib inline"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Do4oeoyNmkj3"
},
"source": [
"### Read Data"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "W_8SVbfwmkj3"
},
"outputs": [],
"source": [
"# Load csv file\n",
"df3 = pd.read_excel('Phase_data.xlsx', na_values=\"na\")"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "miijRwgmmkj4",
"outputId": "20e57ed6-cde0-418b-a511-05538e205794"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1200, 36)"
]
},
"metadata": {},
"execution_count": 4
}
],
"source": [
"# check rows and columns of the dataset\n",
"df3.shape"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 235
},
"id": "ANWg7y2hmkj4",
"outputId": "ac28407e-dec2-40f3-a2a3-b4b7dbf790ec"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" Al Co Cr Fe Ni Cu Mn Ti V Nb ... Pt Y Pd Au \\\n",
"0 0.111 NaN NaN NaN NaN NaN NaN 0.222 0.222 0.222 ... NaN NaN NaN NaN \n",
"1 0.158 NaN NaN NaN NaN NaN NaN 0.215 0.215 0.215 ... NaN NaN NaN NaN \n",
"2 0.588 NaN NaN NaN NaN NaN NaN 0.235 0.235 0.235 ... NaN NaN NaN NaN \n",
"3 0.588 NaN NaN NaN NaN NaN NaN 0.235 0.235 0.235 ... NaN NaN NaN NaN \n",
"4 0.476 NaN NaN NaN NaN NaN NaN 0.239 0.239 NaN ... NaN NaN NaN NaN \n",
"\n",
" dHmix dSmix δ ᐃχ VEC Phases \n",
"0 -8.395 13.146 3.738 0.050 4.556 BCC \n",
"1 -9.352 13.333 3.863 0.233 4.684 BCC \n",
"2 -4.042 12.708 4.003 0.243 4.882 BCC \n",
"3 -4.817 12.708 3.832 0.050 4.647 BCC \n",
"4 -3.356 12.569 4.018 0.244 4.905 BCC \n",
"\n",
"[5 rows x 36 columns]"
],
"text/html": [
"\n",
" <div id=\"df-b44f0d14-2622-433c-a1a7-4e22a96a97f3\">\n",
" <div class=\"colab-df-container\">\n",
" <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>Al</th>\n",
" <th>Co</th>\n",
" <th>Cr</th>\n",
" <th>Fe</th>\n",
" <th>Ni</th>\n",
" <th>Cu</th>\n",
" <th>Mn</th>\n",
" <th>Ti</th>\n",
" <th>V</th>\n",
" <th>Nb</th>\n",
" <th>...</th>\n",
" <th>Pt</th>\n",
" <th>Y</th>\n",
" <th>Pd</th>\n",
" <th>Au</th>\n",
" <th>dHmix</th>\n",
" <th>dSmix</th>\n",
" <th>δ</th>\n",
" <th>ᐃχ</th>\n",
" <th>VEC</th>\n",
" <th>Phases</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>0.111</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.222</td>\n",
" <td>0.222</td>\n",
" <td>0.222</td>\n",
" <td>...</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>-8.395</td>\n",
" <td>13.146</td>\n",
" <td>3.738</td>\n",
" <td>0.050</td>\n",
" <td>4.556</td>\n",
" <td>BCC</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>0.158</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.215</td>\n",
" <td>0.215</td>\n",
" <td>0.215</td>\n",
" <td>...</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>-9.352</td>\n",
" <td>13.333</td>\n",
" <td>3.863</td>\n",
" <td>0.233</td>\n",
" <td>4.684</td>\n",
" <td>BCC</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>0.588</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.235</td>\n",
" <td>0.235</td>\n",
" <td>0.235</td>\n",
" <td>...</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>-4.042</td>\n",
" <td>12.708</td>\n",
" <td>4.003</td>\n",
" <td>0.243</td>\n",
" <td>4.882</td>\n",
" <td>BCC</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>0.588</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.235</td>\n",
" <td>0.235</td>\n",
" <td>0.235</td>\n",
" <td>...</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>-4.817</td>\n",
" <td>12.708</td>\n",
" <td>3.832</td>\n",
" <td>0.050</td>\n",
" <td>4.647</td>\n",
" <td>BCC</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>0.476</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>0.239</td>\n",
" <td>0.239</td>\n",
" <td>NaN</td>\n",
" <td>...</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>NaN</td>\n",
" <td>-3.356</td>\n",
" <td>12.569</td>\n",
" <td>4.018</td>\n",
" <td>0.244</td>\n",
" <td>4.905</td>\n",
" <td>BCC</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>5 rows × 36 columns</p>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-b44f0d14-2622-433c-a1a7-4e22a96a97f3')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-b44f0d14-2622-433c-a1a7-4e22a96a97f3 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-b44f0d14-2622-433c-a1a7-4e22a96a97f3');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
]
},
"metadata": {},
"execution_count": 5
}
],
"source": [
"df3.head()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "EzuTamjymkj4",
"outputId": "d09c6fb6-5d96-413b-ac7d-63ec6f6cae4d"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"MIP 441\n",
"BCC 372\n",
"FCC 220\n",
"FCC_BCC 167\n",
"Name: Phases, dtype: int64"
]
},
"metadata": {},
"execution_count": 6
}
],
"source": [
"# Check unique values of target varaible\n",
"df3['Phases'].value_counts()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "HHD7hYOzmkj5",
"outputId": "20dcf07a-f13b-42bd-d82d-59de761e42fa"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"We have 35 numerical features : ['Al', 'Co', 'Cr', 'Fe', 'Ni', 'Cu', 'Mn', 'Ti', 'V', 'Nb', 'Mo', 'Zr', 'Hf', 'Ta', 'W', 'C', 'Mg', 'Zn', 'Si', 'Re', 'N', 'Li', 'Sn', 'Be', 'B', 'Ag', 'Pt', 'Y', 'Pd', 'Au', 'dHmix', 'dSmix', 'δ', 'ᐃχ', 'VEC']\n",
"\n",
"We have 1 categorical features : ['Phases']\n"
]
}
],
"source": [
"# define numerical & categorical columns\n",
"numeric_features = [feature for feature in df3.columns if df3[feature].dtype != 'O']\n",
"categorical_features = [feature for feature in df3.columns if df3[feature].dtype == 'O']\n",
"\n",
"# print columns\n",
"print('We have {} numerical features : {}'.format(len(numeric_features), numeric_features))\n",
"print('\\nWe have {} categorical features : {}'.format(len(categorical_features), categorical_features))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "_GxWn4Ukmkj6"
},
"source": [
"### Checking missing values"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "8OSwnxNPmkj6",
"outputId": "7adda5ea-14b2-445a-971c-fcc5af1d61ec"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"Al 515\n",
"Co 514\n",
"Cr 442\n",
"Fe 386\n",
"Ni 378\n",
"Cu 858\n",
"Mn 971\n",
"Ti 754\n",
"V 954\n",
"Nb 882\n",
"Mo 969\n",
"Zr 964\n",
"Hf 1081\n",
"Ta 1074\n",
"W 1164\n",
"C 1158\n",
"Mg 1161\n",
"Zn 1165\n",
"Si 1137\n",
"Re 1199\n",
"N 1194\n",
"Li 1179\n",
"Sn 1160\n",
"Be 1199\n",
"B 1196\n",
"Ag 1199\n",
"Pt 1199\n",
"Y 1194\n",
"Pd 1198\n",
"Au 1199\n",
"dHmix 0\n",
"dSmix 0\n",
"δ 0\n",
"ᐃχ 0\n",
"VEC 0\n",
"Phases 0\n",
"dtype: int64"
]
},
"metadata": {},
"execution_count": 8
}
],
"source": [
"df3.isnull().sum()"
]
},
{
"cell_type": "code",
"source": [
"feature_names = ['Al', 'Co', 'Cr', 'Fe', 'Ni', 'Cu', 'Mn', 'Ti', 'V', 'Nb', 'Mo', 'Zr',\n",
" 'Hf', 'Ta', 'W', 'C', 'Mg', 'Zn', 'Si', 'Re', 'N', 'Li', 'Sn', 'Be',\n",
" 'B', 'Ag', 'Pt', 'Y', 'Pd', 'Au', 'dHmix', 'dSmix', 'δ',\n",
" 'ᐃχ', 'VEC', 'Phases']"
],
"metadata": {
"id": "Bqb2WLli8aio"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# sns.catplot(data=df33, x=\"Phases\", kind=\"count\", palette=\"winter_r\")\n",
"# plt.show()\n",
"# sns.set_style('whitegrid')\n",
"# sns.set(font_scale = 1.2)\n",
"\n",
"# df3['Phases'].value_counts()\n",
"\n",
"# p= sns.countplot(df3['Phases'], dodge=False , hatch='/',palette=\"Accent\", ec='black')\n",
"# # sns.set(rc={'figure.figsize': (8,10)})\n",
"\n",
"# p.set_xlabel(\"Phases\")\n",
"# p.set_ylabel(\"Samples\")"
],
"metadata": {
"id": "lv0NNLuwynDH"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"df3['Phases'].replace({'MIP':0, 'BCC':1, 'FCC':2, 'FCC_BCC':3}, inplace=True) "
],
"metadata": {
"id": "gana74hF9F6V"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"from sklearn.impute import KNNImputer\n",
"from sklearn.pipeline import Pipeline\n",
"imputer = KNNImputer(n_neighbors=10)\n",
"df_imp = imputer.fit_transform(df3)\n",
"df_afterimp =pd.DataFrame(df_imp, columns=feature_names[0:36])"
],
"metadata": {
"id": "-Mm5xHNY8Eze"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"source": [
"outlier removal after missing value imputation\n"
],
"metadata": {
"id": "yNVhMH0m_hu-"
}
},
{
"cell_type": "code",
"source": [
"df = df_afterimp.copy()"
],
"metadata": {
"id": "HkRVz8WW_n_U"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"def IQR_capping(df_afterimp, cols, factor):\n",
" \n",
" for col in cols:\n",
" Q1 = df_afterimp[col].quantile(0.10)\n",
" Q3 = df_afterimp[col].quantile(0.90)\n",
" IQR=Q3-Q1\n",
"\n",
" lower_boundary = Q1-(factor*IQR)\n",
" upper_boundary = Q3+(factor*IQR)\n",
" \n",
"\n",
" df_afterimp[col] = np.where(df_afterimp[col]>upper_boundary, upper_boundary, np.where(df_afterimp[col]<lower_boundary, lower_boundary, df_afterimp[col]))"
],
"metadata": {
"id": "mS450J8m_ux1"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"IQR_capping(df, feature_names, 1.5)"
],
"metadata": {
"id": "eS3oxy9-_v3r"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# for col in feature_names:\n",
"# plt.figure(figsize=(16,4))\n",
"\n",
"# plt.subplot(141)\n",
"# sns.distplot(df_afterimp[col], label='skew: '+ str(np.round(df_afterimp[col].skew(),2)))\n",
"# plt.title('Before')\n",
"# plt.legend()\n",
"\n",
"# plt.subplot(142)\n",
"# sns.distplot(df[col], label ='skew '+ str(np.round(df[col].skew(),2)))\n",
"# plt.title('After')\n",
"# plt.legend()\n",
"\n",
"# plt.subplot(143)\n",
"# sns.boxplot(df_afterimp[col])\n",
"# plt.title('Before')\n",
"\n",
"# plt.subplot(144)\n",
"# sns.boxplot(df[col])\n",
"# plt.title('After')\n",
"# plt.tight_layout()\n",
"# plt.show()"
],
"metadata": {
"id": "CfPevh6E_y8s"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "e3SW6fs3mkj8"
},
"source": [
"## Visualization of unique values in Target variable"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "pRzK0OL9mkj9"
},
"outputs": [],
"source": [
"## Visualization of unique values in Target variable\n",
"# sns.catplot(data=df, x=\"Phases\", kind=\"count\", palette=\"winter_r\", alpha=.6)\n",
"# plt.show()\n"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "N5RYFNGjmkj9"
},
"source": [
"## Create Functions for model training and evaluation"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "snr8L-EVmkj9"
},
"outputs": [],
"source": [
"def evaluate_clf(true, predicted):\n",
" '''\n",
" This function takes in true values and predicted values\n",
" Returns: Accuracy, F1-Score, Precision, Recall\n",
" '''\n",
" acc = accuracy_score(true, predicted) # Calculate Accuracy\n",
" f1 = f1_score(true, predicted,average='weighted') # Calculate F1-score\n",
" precision = precision_score(true, predicted, average='weighted') # Calculate Precision\n",
" recall = recall_score(true, predicted, average='weighted') # Calculate Recall\n",
" \n",
" return acc, f1 , precision, recall"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "qSI82B3xmkj-"
},
"outputs": [],
"source": [
"# Create a function which can evaluate models and return a report \n",
"def evaluate_models(X, y, models):\n",
" '''\n",
" This function takes in X and y and models dictionary as input\n",
" It splits the data into Train Test split\n",
" Iterates through the given model dictionary and evaluates the metrics\n",
" Returns: Dataframe which contains report of all models metrics\n",
" '''\n",
" # separate dataset into train and test\n",
" X_train, X_test, y_train, y_test = train_test_split(X,y,test_size=0.2,random_state=42)\n",
" \n",
" cost_list=[]\n",
" models_list = []\n",
" accuracy_list = []\n",
" \n",
" for i in range(len(list(models))):\n",
" model = list(models.values())[i]\n",
" model.fit(X_train, y_train) # Train model\n",
"\n",
" # Make predictions\n",
" y_train_pred = model.predict(X_train)\n",
" y_test_pred = model.predict(X_test)\n",
"\n",
" # Training set performance\n",
" model_train_accuracy, model_train_f1,model_train_precision,\\\n",
" model_train_recall=evaluate_clf(y_train ,y_train_pred)\n",
"\n",
"\n",
" # Test set performance\n",
" model_test_accuracy,model_test_f1,model_test_precision,\\\n",
" model_test_recall=evaluate_clf(y_test, y_test_pred)\n",
" \n",
"\n",
" print(list(models.keys())[i])\n",
" models_list.append(list(models.keys())[i])\n",
"\n",
" print('Model performance for Training set')\n",
" print(\"- Accuracy: {:.4f}\".format(model_train_accuracy))\n",
" print('- F1 score: {:.4f}'.format(model_train_f1)) \n",
" print('- Precision: {:.4f}'.format(model_train_precision))\n",
" print('- Recall: {:.4f}'.format(model_train_recall))\n",
" \n",
"\n",
" print('----------------------------------')\n",
"\n",
" print('Model performance for Test set')\n",
" print('- Accuracy: {:.4f}'.format(model_test_accuracy))\n",
" print('- F1 score: {:.4f}'.format(model_test_f1))\n",
" print('- Precision: {:.4f}'.format(model_test_precision))\n",
" print('- Recall: {:.4f}'.format(model_test_recall))\n",
" \n",
" print('='*35)\n",
" print('\\n')\n",
" \n",
" report=pd.DataFrame(list(zip(models_list, cost_list)), columns=['Model Name', 'Cost']).sort_values(by=[\"Cost\"])\n",
" \n",
" return report"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "5k1CaW9cmkj-"
},
"source": [
"### Plot distribution of all Independent Numerical variables"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "YvJ2TGK9mkj-"
},
"outputs": [],
"source": [
"# numeric_features = [feature for feature in df.columns if df[feature].dtype != 'O']\n",
"\n",
"# plt.figure(figsize=(15, 100))\n",
"# for i, col in enumerate(numeric_features):\n",
"# plt.subplot(60, 3, i+1)\n",
"# sns.distplot(x=df[col], color='indianred')\n",
"# plt.xlabel(col, weight='bold')\n",
"# plt.tight_layout()"
]
},
{
"cell_type": "code",
"source": [
"df2= df.copy()\n",
"Al_trans = np.log(df2['Al'])\n",
"Co_trans = np.log(df2['Co'])\n",
"Cr_trans = np.log(df2['Cr'])\n",
"Fe_trans = np.log(df2['Fe'])\n",
"Ni_trans = np.log(df2['Ni'])\n",
"Cu_trans = np.log(df2['Cu'])\n",
"Mn_trans = np.log(df2['Mn'])\n",
"Ti_trans = np.log(df2['Ti'])\n",
"V_trans = np.log(df2['V'])\n",
"Nb_trans = np.log(df2['Nb'])\n",
"Mo_trans = np.log(df2['Mo'])\n",
"Zr_trans = np.log(df2['Zr'])\n",
"Hf_trans = np.log(df2['Hf'])\n",
"Ta_trans = np.log(df2['Ta'])\n",
"W_trans = np.log(df2['W'])\n",
"C_trans = np.log(df2['C'])\n",
"Mg_trans = np.log(df2['Mg'])\n",
"Zn_trans = np.log(df2['Zn'])\n",
"Si_trans = np.log(df2['Si'])\n",
"Re_trans = np.log(df2['Re'])\n",
"N_trans = np.log(df2['N'])\n",
"Li_trans = np.log(df2['Li'])\n",
"Sn_trans = np.log(df2['Sn'])\n",
"Be_trans = np.log(df2['Sn'])\n",
"B_trans = np.log(df2['B'])\n",
"Ag_trans = np.log(df2['Ag'])\n",
"Pt_trans = np.log(df2['Pt'])\n",
"Y_trans = np.log(df2['Y'])\n",
"Pd_trans = np.log(df2['Pd'])\n",
"Au_trans = np.log(df2['Au'])\n",
"#dHmix_trans = np.log(df2['dHmix'])\n",
"dSmix_trans = np.log(df2['dSmix'])\n",
"Atom_Size_Diff_trans = np.log(df2['δ'])\n",
"Elect_Diff_trans = np.log(df2['ᐃχ'])\n",
"VEC_trans = np.log(df2['VEC']) "
],
"metadata": {
"id": "xhftSkwu2TuH"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"df_final = pd.DataFrame(pd.concat([Al_trans, Co_trans, Cr_trans, Fe_trans, Ni_trans, Cu_trans, Mn_trans, \n",
"Ti_trans, V_trans,Nb_trans, Mo_trans, Zr_trans, Hf_trans, Ta_trans, W_trans, C_trans, Mg_trans, Zn_trans, Si_trans, Re_trans, N_trans, \n",
"Li_trans, Sn_trans, Be_trans, B_trans, Ag_trans, Pt_trans, Y_trans, Pd_trans, Au_trans, df2['dHmix'], dSmix_trans, Atom_Size_Diff_trans, \n",
"Elect_Diff_trans, VEC_trans, df2['Phases']], axis=1))\n",
"df_final"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 487
},
"id": "YqyteeKkBzb2",
"outputId": "3681e51b-f3c8-4f22-ee69-a4851d9b19df"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" Al Co Cr Fe Ni Cu Mn \\\n",
"0 -2.198225 -1.445195 -0.998586 -1.275111 -1.738977 -1.192044 -1.534794 \n",
"1 -1.845160 -1.582796 -1.289894 -1.427116 -1.581823 -1.543650 -1.541312 \n",
"2 -0.531028 -1.595535 -1.283377 -1.352473 -1.621005 -0.931404 -1.552113 \n",
"3 -0.531028 -1.581336 -1.375552 -1.322381 -1.565421 -1.119325 -1.539446 \n",
"4 -0.742337 -1.716466 -1.388296 -1.358679 -1.543650 -0.886004 -1.566378 \n",
"... ... ... ... ... ... ... ... \n",
"1195 -1.944911 -1.944911 -1.944911 -1.944911 -1.944911 -1.944911 -1.308223 \n",
"1196 -1.697723 -1.789761 -1.789761 -1.789761 -1.789761 -0.813509 -1.789761 \n",
"1197 -1.002666 -1.789761 -1.789761 -1.789761 -1.789761 -1.085301 -1.789761 \n",
"1198 -1.128247 -1.944911 -1.944911 -1.944911 -1.944911 -1.944911 -1.810942 \n",
"1199 -0.877070 -1.609438 -1.505979 -1.609438 -0.916291 -1.488106 -1.609438 \n",
"\n",
" Ti V Nb ... Pt Y Pd \\\n",
"0 -1.505078 -1.505078 -1.505078 ... -1.609438 -1.806873 -1.203973 \n",
"1 -1.537117 -1.537117 -1.537117 ... -1.609438 -1.806873 -1.203973 \n",
"2 -1.448170 -1.448170 -1.448170 ... -1.609438 -1.806873 -1.203973 \n",
"3 -1.448170 -1.448170 -1.448170 ... -1.609438 -1.806873 -1.203973 \n",
"4 -1.431292 -1.431292 -1.480166 ... -1.609438 -1.806873 -1.203973 \n",
"... ... ... ... ... ... ... ... \n",
"1195 -0.604770 -1.061895 -0.568631 ... -1.609438 -1.806873 -1.203973 \n",
"1196 -1.864330 -1.101115 -1.789761 ... -1.609438 -1.806873 -1.203973 \n",
"1197 -1.789761 -1.113827 -1.515948 ... -1.609438 -1.806873 -1.203973 \n",
"1198 -1.251763 -1.013077 -1.197659 ... -1.609438 -1.806873 -1.203973 \n",
"1199 -0.909315 -1.855978 -0.944176 ... -1.609438 -1.806873 -1.203973 \n",
"\n",
" Au dHmix dSmix δ ᐃχ VEC Phases \n",
"0 -1.789761 -8.395 2.576118 1.318551 -2.995732 1.516445 1.0 \n",
"1 -1.789761 -9.352 2.590242 1.351444 -1.456717 1.544152 1.0 \n",
"2 -1.789761 -4.042 2.542232 1.387044 -1.414694 1.585555 1.0 \n",
"3 -1.789761 -4.817 2.542232 1.343387 -2.995732 1.536222 1.0 \n",
"4 -1.789761 -3.356 2.531233 1.390784 -1.410587 1.590255 1.0 \n",
"... ... ... ... ... ... ... ... \n",
"1195 -1.789761 -18.857 2.783467 1.812542 -2.154165 1.985955 0.0 \n",
"1196 -1.789761 -12.000 2.700958 1.703839 -1.958995 2.014903 0.0 \n",
"1197 -1.789761 -13.444 2.700958 1.839438 -1.883875 1.992385 0.0 \n",
"1198 -1.789761 -14.041 2.676147 1.979759 -1.864330 2.005391 0.0 \n",
"1199 -1.789761 -4.160 2.404239 1.190888 -1.994365 2.174752 2.0 \n",
"\n",
"[1200 rows x 36 columns]"
],
"text/html": [
"\n",
" <div id=\"df-c73e2d5d-6b28-46f1-ab87-71c17cc63a90\">\n",
" <div class=\"colab-df-container\">\n",
" <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>Al</th>\n",
" <th>Co</th>\n",
" <th>Cr</th>\n",
" <th>Fe</th>\n",
" <th>Ni</th>\n",
" <th>Cu</th>\n",
" <th>Mn</th>\n",
" <th>Ti</th>\n",
" <th>V</th>\n",
" <th>Nb</th>\n",
" <th>...</th>\n",
" <th>Pt</th>\n",
" <th>Y</th>\n",
" <th>Pd</th>\n",
" <th>Au</th>\n",
" <th>dHmix</th>\n",
" <th>dSmix</th>\n",
" <th>δ</th>\n",
" <th>ᐃχ</th>\n",
" <th>VEC</th>\n",
" <th>Phases</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>-2.198225</td>\n",
" <td>-1.445195</td>\n",
" <td>-0.998586</td>\n",
" <td>-1.275111</td>\n",
" <td>-1.738977</td>\n",
" <td>-1.192044</td>\n",
" <td>-1.534794</td>\n",
" <td>-1.505078</td>\n",
" <td>-1.505078</td>\n",
" <td>-1.505078</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-8.395</td>\n",
" <td>2.576118</td>\n",
" <td>1.318551</td>\n",
" <td>-2.995732</td>\n",
" <td>1.516445</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>-1.845160</td>\n",
" <td>-1.582796</td>\n",
" <td>-1.289894</td>\n",
" <td>-1.427116</td>\n",
" <td>-1.581823</td>\n",
" <td>-1.543650</td>\n",
" <td>-1.541312</td>\n",
" <td>-1.537117</td>\n",
" <td>-1.537117</td>\n",
" <td>-1.537117</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-9.352</td>\n",
" <td>2.590242</td>\n",
" <td>1.351444</td>\n",
" <td>-1.456717</td>\n",
" <td>1.544152</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>-0.531028</td>\n",
" <td>-1.595535</td>\n",
" <td>-1.283377</td>\n",
" <td>-1.352473</td>\n",
" <td>-1.621005</td>\n",
" <td>-0.931404</td>\n",
" <td>-1.552113</td>\n",
" <td>-1.448170</td>\n",
" <td>-1.448170</td>\n",
" <td>-1.448170</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-4.042</td>\n",
" <td>2.542232</td>\n",
" <td>1.387044</td>\n",
" <td>-1.414694</td>\n",
" <td>1.585555</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>-0.531028</td>\n",
" <td>-1.581336</td>\n",
" <td>-1.375552</td>\n",
" <td>-1.322381</td>\n",
" <td>-1.565421</td>\n",
" <td>-1.119325</td>\n",
" <td>-1.539446</td>\n",
" <td>-1.448170</td>\n",
" <td>-1.448170</td>\n",
" <td>-1.448170</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-4.817</td>\n",
" <td>2.542232</td>\n",
" <td>1.343387</td>\n",
" <td>-2.995732</td>\n",
" <td>1.536222</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>-0.742337</td>\n",
" <td>-1.716466</td>\n",
" <td>-1.388296</td>\n",
" <td>-1.358679</td>\n",
" <td>-1.543650</td>\n",
" <td>-0.886004</td>\n",
" <td>-1.566378</td>\n",
" <td>-1.431292</td>\n",
" <td>-1.431292</td>\n",
" <td>-1.480166</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-3.356</td>\n",
" <td>2.531233</td>\n",
" <td>1.390784</td>\n",
" <td>-1.410587</td>\n",
" <td>1.590255</td>\n",
" <td>1.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>...</th>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" <td>...</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1195</th>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.308223</td>\n",
" <td>-0.604770</td>\n",
" <td>-1.061895</td>\n",
" <td>-0.568631</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-18.857</td>\n",
" <td>2.783467</td>\n",
" <td>1.812542</td>\n",
" <td>-2.154165</td>\n",
" <td>1.985955</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1196</th>\n",
" <td>-1.697723</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-0.813509</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.864330</td>\n",
" <td>-1.101115</td>\n",
" <td>-1.789761</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-12.000</td>\n",
" <td>2.700958</td>\n",
" <td>1.703839</td>\n",
" <td>-1.958995</td>\n",
" <td>2.014903</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1197</th>\n",
" <td>-1.002666</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.085301</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.789761</td>\n",
" <td>-1.113827</td>\n",
" <td>-1.515948</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-13.444</td>\n",
" <td>2.700958</td>\n",
" <td>1.839438</td>\n",
" <td>-1.883875</td>\n",
" <td>1.992385</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1198</th>\n",
" <td>-1.128247</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.944911</td>\n",
" <td>-1.810942</td>\n",
" <td>-1.251763</td>\n",
" <td>-1.013077</td>\n",
" <td>-1.197659</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-14.041</td>\n",
" <td>2.676147</td>\n",
" <td>1.979759</td>\n",
" <td>-1.864330</td>\n",
" <td>2.005391</td>\n",
" <td>0.0</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1199</th>\n",
" <td>-0.877070</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.505979</td>\n",
" <td>-1.609438</td>\n",
" <td>-0.916291</td>\n",
" <td>-1.488106</td>\n",
" <td>-1.609438</td>\n",
" <td>-0.909315</td>\n",
" <td>-1.855978</td>\n",
" <td>-0.944176</td>\n",
" <td>...</td>\n",
" <td>-1.609438</td>\n",
" <td>-1.806873</td>\n",
" <td>-1.203973</td>\n",
" <td>-1.789761</td>\n",
" <td>-4.160</td>\n",
" <td>2.404239</td>\n",
" <td>1.190888</td>\n",
" <td>-1.994365</td>\n",
" <td>2.174752</td>\n",
" <td>2.0</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"<p>1200 rows × 36 columns</p>\n",
"</div>\n",
" <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-c73e2d5d-6b28-46f1-ab87-71c17cc63a90')\"\n",
" title=\"Convert this dataframe to an interactive table.\"\n",
" style=\"display:none;\">\n",
" \n",
" <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
" width=\"24px\">\n",
" <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
" <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
" </svg>\n",
" </button>\n",
" \n",
" <style>\n",
" .colab-df-container {\n",
" display:flex;\n",
" flex-wrap:wrap;\n",
" gap: 12px;\n",
" }\n",
"\n",
" .colab-df-convert {\n",
" background-color: #E8F0FE;\n",
" border: none;\n",
" border-radius: 50%;\n",
" cursor: pointer;\n",
" display: none;\n",
" fill: #1967D2;\n",
" height: 32px;\n",
" padding: 0 0 0 0;\n",
" width: 32px;\n",
" }\n",
"\n",
" .colab-df-convert:hover {\n",
" background-color: #E2EBFA;\n",
" box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
" fill: #174EA6;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert {\n",
" background-color: #3B4455;\n",
" fill: #D2E3FC;\n",
" }\n",
"\n",
" [theme=dark] .colab-df-convert:hover {\n",
" background-color: #434B5C;\n",
" box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
" filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
" fill: #FFFFFF;\n",
" }\n",
" </style>\n",
"\n",
" <script>\n",
" const buttonEl =\n",
" document.querySelector('#df-c73e2d5d-6b28-46f1-ab87-71c17cc63a90 button.colab-df-convert');\n",
" buttonEl.style.display =\n",
" google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
"\n",
" async function convertToInteractive(key) {\n",
" const element = document.querySelector('#df-c73e2d5d-6b28-46f1-ab87-71c17cc63a90');\n",
" const dataTable =\n",
" await google.colab.kernel.invokeFunction('convertToInteractive',\n",
" [key], {});\n",
" if (!dataTable) return;\n",
"\n",
" const docLinkHtml = 'Like what you see? Visit the ' +\n",
" '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
" + ' to learn more about interactive tables.';\n",
" element.innerHTML = '';\n",
" dataTable['output_type'] = 'display_data';\n",
" await google.colab.output.renderOutput(dataTable, element);\n",
" const docLink = document.createElement('div');\n",
" docLink.innerHTML = docLinkHtml;\n",
" element.appendChild(docLink);\n",
" }\n",
" </script>\n",
" </div>\n",
" </div>\n",
" "
]
},
"metadata": {},
"execution_count": 911
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "IcaPyF-cmkj-"
},
"source": [
"# Evaluate Model on Different experiments"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "ZLnlk8lLmkj_"
},
"outputs": [],
"source": [
"# Splitting X and y for all Experiments\n",
"X= df_final.drop('Phases', axis=1)\n",
"y = df_final['Phases']"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "19sBusUMmkj_"
},
"outputs": [],
"source": [
"y= y.replace({'MIP': 0, 'BCC': 1,'FCC': 2,'FCC_BCC': 3})"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "gYETL8wmmkj_"
},
"source": [
"#### All datset is not normally distributed so I am going for robust scaler, if you want you can do the log trabnsfrmation and then standard scaler to see the result"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "AGRi1lDdmkj_"
},
"outputs": [],
"source": [
"# Fit with robust scaler for KNN best K-selection experminet\n",
"robustscaler = RobustScaler()\n",
"X1 = robustscaler.fit_transform(X)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "K5UXXvl0mkj_"
},
"outputs": [],
"source": [
"results=[]\n",
"# define imputer\n",
"imputer = KNNImputer(n_neighbors=5, weights='uniform', metric='nan_euclidean')\n",
"strategies = [str(i) for i in [1,3,5,7,9]]\n",
"for s in strategies:\n",
" pipeline = Pipeline(steps=[('i', KNNImputer(n_neighbors=int(s))), ('m', LogisticRegression())])\n",
" scores = cross_val_score(pipeline, X1, y, scoring='accuracy', cv=2, n_jobs=-1)\n",
" results.append(scores)\n",
" #print('n_neighbors= %s || accuracy (%.4f)' % (s , mean(scores)))"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "Lc9ydKsbmkj_"
},
"outputs": [],
"source": [
"### Pipeline for KNN imputer\n",
"num_features = X.select_dtypes(exclude=\"object\").columns\n",
"\n",
"# Fit the KNN imputer with selected K-value\n",
"knn_pipeline = Pipeline(steps=[\n",
" ('imputer', KNNImputer(n_neighbors=3)),\n",
" ('RobustScaler', RobustScaler())\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "LNeqM7AqmkkA"
},
"outputs": [],
"source": [
"X_knn =knn_pipeline.fit_transform(X)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "U2WtW76UmkkA"
},
"source": [
"## Handling Imbalanced data using SMOTE-Tomek links"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "a_qoKqBImkkA"
},
"outputs": [],
"source": [
"from imblearn.combine import SMOTETomek\n",
"\n",
"# Resampling the minority class. The strategy can be changed as required.\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority',n_jobs=-1)\n",
"# Fit the model to generate the data.\n",
"X_res, y_res = smt.fit_resample(X_knn, y)"
]
},
{
"cell_type": "markdown",
"source": [
"##### Number of samples for SMOTE-Tomek link data"
],
"metadata": {
"id": "Jv2WfROa_Jqf"
}
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "gnCJnOrimkkA",
"outputId": "ea6b533c-fee9-4143-eb1e-b917ca41588f"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1392,)"
]
},
"metadata": {},
"execution_count": 919
}
],
"source": [
"y_res.shape "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "RJSirT9GmkkA",
"outputId": "98364b68-8b33-4978-b8ea-2a290d30063f"
},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1392, 35)"
]
},
"metadata": {},
"execution_count": 920
}
],
"source": [
"X_res.shape"
]
},
{
"cell_type": "markdown",
"source": [
"##### Visualization of samples for SMOTE-Tomek link data"
],
"metadata": {
"id": "gTSW-AQX_Auv"
}
},
{
"cell_type": "code",
"source": [
"# ## Visualization of unique values in Target variable\n",
"# sns.set_style('whitegrid')\n",
"# sns.set(font_scale = 1.2)\n",
"# y_res= y_res.replace({0:'MIP', 1:'BCC', 2:'FCC', 3:'FCC_BCC'})\n",
"# y_res.value_counts()\n",
"\n",
"# p= sns.countplot(y_res, dodge=False,palette=\"Accent\", ec='black')\n",
"# # sns.set(rc={'figure.figsize': (8,10)})\n",
"# p.set_xlabel(\"Phases\")\n",
"# p.set_ylabel(\"Samples\")"
],
"metadata": {
"id": "pycJqRHdxpNl"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "mJqwtYDWmkkB"
},
"source": [
"### Initialize Default Models in a dictionary"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "nz1HT0QAmkkB"
},
"outputs": [],
"source": [
"# Dictionary which contains models for experiment\n",
"models = {\n",
" \"Random Forest\": RandomForestClassifier(),\n",
" \"Decision Tree\": DecisionTreeClassifier(),\n",
" \"K-Neighbors Classifier\": KNeighborsClassifier(),\n",
" \"XGBClassifier\": XGBClassifier(), \n",
" \"SVM\": SVC()\n",
"}"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "xy77_DslmkkB"
},
"source": [
"### Fit KNN imputed data for models in dictionary"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "6eJUzP0TmkkB",
"outputId": "4f189dbb-5705-4eeb-a611-bf42545ce187"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Random Forest\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.9211\n",
"- F1 score: 0.9209\n",
"- Precision: 0.9216\n",
"- Recall: 0.9211\n",
"===================================\n",
"\n",
"\n",
"Decision Tree\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8459\n",
"- F1 score: 0.8461\n",
"- Precision: 0.8471\n",
"- Recall: 0.8459\n",
"===================================\n",
"\n",
"\n",
"K-Neighbors Classifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.8302\n",
"- F1 score: 0.8284\n",
"- Precision: 0.8325\n",
"- Recall: 0.8302\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.7957\n",
"- F1 score: 0.7910\n",
"- Precision: 0.7970\n",
"- Recall: 0.7957\n",
"===================================\n",
"\n",
"\n",
"XGBClassifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.9443\n",
"- F1 score: 0.9440\n",
"- Precision: 0.9442\n",
"- Recall: 0.9443\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8925\n",
"- F1 score: 0.8930\n",
"- Precision: 0.8971\n",
"- Recall: 0.8925\n",
"===================================\n",
"\n",
"\n",
"SVM\n",
"Model performance for Training set\n",
"- Accuracy: 0.8392\n",
"- F1 score: 0.8370\n",
"- Precision: 0.8419\n",
"- Recall: 0.8392\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8208\n",
"- F1 score: 0.8164\n",
"- Precision: 0.8293\n",
"- Recall: 0.8208\n",
"===================================\n",
"\n",
"\n"
]
}
],
"source": [
"report_knn = evaluate_models(X_res, y_res, models)"
]
},
{
"cell_type": "code",
"source": [
"full_data= df_final.copy()\n",
"X_full= full_data.drop('Phases', axis=1)\n",
"y_full = full_data['Phases']"
],
"metadata": {
"id": "_kgNC0ANuQeL"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"knn_pipeline2= Pipeline(steps=[('imputer', KNNImputer(n_neighbors=3)),('RobustScaler', RobustScaler())])\n",
"X_knn2 =knn_pipeline2.fit_transform(X_full)\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority',n_jobs=-1)\n",
"# Fit the model to generate the data.\n",
"X_res, y_res = smt.fit_resample(X_knn2, y_full)\n",
"X_train_fullkn, X_test_fullkn, y_train_fullkn, y_test_fullkn = train_test_split(X_res, y_res, test_size=0.20, random_state=42)\n",
"\n",
"## Hyper-tuned RFC\n",
"model_rf_hyperkn = RandomForestClassifier( ).fit(X_train_fullkn, y_train_fullkn)\n",
"probs_rf_hyper = model_rf_hyperkn .predict_proba(X_test_fullkn)#[:, 1]\n",
"\n",
"auc_rf = roc_auc_score(y_test_fullkn, probs_rf_hyper, multi_class='ovr')\n",
"print('AUC of RFC: {:.4f}'.format(auc_rf))\n",
"\n",
"\n",
"RFC_KNN = RandomForestClassifier() #random_state= 100, min_samples_split= 6,\n",
" \n",
"scores_cv_rf= cross_val_score(RFC_KNN, X_knn2, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores_cv_rf_std= cross_val_score(RFC_KNN, X_knn2, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"#print('10-fold CV of RFC', (scores_cv_rf))\n",
"print('10-fold CV mean of RFC: {:.4f}'.format(scores_cv_rf))\n",
"print('10-fold CV std of RFC: {:.4f}'.format(scores_cv_rf_std))"
],
"metadata": {
"id": "kIlLH9Qnt9AG",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "c8e95b28-ffb3-4a49-c2d0-c9794b382901"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"AUC of RFC: 0.9841\n",
"10-fold CV mean of RFC: 0.9262\n",
"10-fold CV std of RFC: 0.0307\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"print(X_train_fullkn.shape)\n",
"print(X_test_fullkn.shape)\n",
"print(y_train_fullkn.shape)\n",
"print(y_test_fullkn.shape)"
],
"metadata": {
"id": "b4T_WjV6z5uy",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "0a307dd9-12d6-4956-8940-0c570eac97b8"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"(1113, 35)\n",
"(279, 35)\n",
"(1113,)\n",
"(279,)\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"y_test_fullkn.value_counts()"
],
"metadata": {
"id": "b5em3eaK1IBv",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "d660d86f-e630-4586-d447-343293bb5e10"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"3.0 94\n",
"0.0 78\n",
"1.0 67\n",
"2.0 40\n",
"Name: Phases, dtype: int64"
]
},
"metadata": {},
"execution_count": 928
}
]
},
{
"cell_type": "code",
"source": [
"\n",
"DTC = DecisionTreeClassifier()\n",
"RFC = RandomForestClassifier()\n",
"XGB = XGBClassifier()\n",
"KNN = KNeighborsClassifier()\n",
"SVM = SVC(probability=True)\n",
"\n",
"#scores1_cv_cb= cross_val_score(CBC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores1_cv_dt= cross_val_score(DTC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores1_cv_rf= cross_val_score(RFC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores1_cv_xg= cross_val_score(XGB, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores1_cv_kn= cross_val_score(KNN, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"scores1_cv_SV= cross_val_score(SVM, X_knn, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"\n",
"#scores1_std_cb= cross_val_score(CBC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"scores1_std_dt= cross_val_score(DTC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"scores1_std_rf= cross_val_score(RFC, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"scores1_std_xg= cross_val_score(XGB, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"scores1_std_kn= cross_val_score(KNN, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"scores1_std_SV= cross_val_score(SVM, X_knn, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"\n",
"#print('10-fold CV for CBC', (scores1_cv_cb)) \n",
"print('10-fold CV forDTC', (scores1_cv_dt)) \n",
"print('10-fold CV forRFC', (scores1_cv_rf))\n",
"print('10-fold CV forXGB',(scores1_cv_xg))\n",
"print('10-fold CV forKNN',(scores1_cv_kn))\n",
"print('10-fold CV forSVM',(scores1_cv_SV))\n",
"\n",
"#print('std for CBC', (scores1_std_cb)) \n",
"print('std forDTC', (scores1_std_dt)) \n",
"print('std for RFC', (scores1_std_rf))\n",
"print('std for XGB',(scores1_std_xg))\n",
"print('std for KNN',(scores1_std_kn))\n",
"print('std for SVM',(scores1_std_SV))"
],
"metadata": {
"id": "YT423e-hupzS",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "bf62567f-19af-41d2-8a3a-2d0a665e0c7f"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"10-fold CV forDTC 0.7667898532261462\n",
"10-fold CV forRFC 0.9211907342671587\n",
"10-fold CV forXGB 0.905462515029783\n",
"10-fold CV forKNN 0.8233527212676955\n",
"10-fold CV forSVM 0.8677643965852303\n",
"std forDTC 0.04311219027196229\n",
"std for RFC 0.02890207696361916\n",
"std for XGB 0.03213222959377166\n",
"std for KNN 0.05565549213845232\n",
"std for SVM 0.057071330411747924\n"
]
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "yjk_MgxRmkkB"
},
"source": [
"### Experiment: 2 = Simple Imputer with Strategy Median "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "X7iqoWQUmkkB"
},
"outputs": [],
"source": [
"num_features = X.select_dtypes(exclude=\"object\").columns\n",
"\n",
"# Fit the Simple imputer with strategy median\n",
"median_pipeline = Pipeline(steps=[\n",
" ('imputer', SimpleImputer(strategy='median')),\n",
" ('RobustScaler', RobustScaler())\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "M7q92I-WmkkB"
},
"outputs": [],
"source": [
"# Fit X with median_pipeline\n",
"X_median = median_pipeline.fit_transform(X)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "EGy0ilAxmkkC"
},
"outputs": [],
"source": [
"# Resampling the minority class. The strategy can be changed as required.\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority')\n",
"# Fit the model to generate the data.\n",
"X_res1, y_res1 = smt.fit_resample(X_median, y)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "NfbuhaBSmkkC",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "16738bb8-cd01-4254-9e55-1a98ccaa3015"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Random Forest\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.9140\n",
"- F1 score: 0.9143\n",
"- Precision: 0.9154\n",
"- Recall: 0.9140\n",
"===================================\n",
"\n",
"\n",
"Decision Tree\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8351\n",
"- F1 score: 0.8354\n",
"- Precision: 0.8363\n",
"- Recall: 0.8351\n",
"===================================\n",
"\n",
"\n",
"K-Neighbors Classifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.8293\n",
"- F1 score: 0.8277\n",
"- Precision: 0.8282\n",
"- Recall: 0.8293\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8029\n",
"- F1 score: 0.8000\n",
"- Precision: 0.8033\n",
"- Recall: 0.8029\n",
"===================================\n",
"\n",
"\n",
"XGBClassifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.9443\n",
"- F1 score: 0.9440\n",
"- Precision: 0.9442\n",
"- Recall: 0.9443\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8925\n",
"- F1 score: 0.8930\n",
"- Precision: 0.8971\n",
"- Recall: 0.8925\n",
"===================================\n",
"\n",
"\n",
"SVM\n",
"Model performance for Training set\n",
"- Accuracy: 0.8392\n",
"- F1 score: 0.8371\n",
"- Precision: 0.8421\n",
"- Recall: 0.8392\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8208\n",
"- F1 score: 0.8164\n",
"- Precision: 0.8293\n",
"- Recall: 0.8208\n",
"===================================\n",
"\n",
"\n"
]
}
],
"source": [
"# Training the models\n",
"report_median = evaluate_models(X_res1, y_res1, models)"
]
},
{
"cell_type": "code",
"source": [
"median_pipeline1 = Pipeline(steps=[\n",
" ('imputer', SimpleImputer(strategy='median')),\n",
" ('RobustScaler', RobustScaler())\n",
"])\n",
"X_median_full = median_pipeline1.fit_transform(X_full)\n",
"\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority',n_jobs=-1)\n",
"# Fit the model to generate the data.\n",
"X_res2, y_res2 = smt.fit_resample(X_median_full , y_full)\n",
"X_train_full_median, X_test_full_median, y_train_full_median, y_test_full_median = train_test_split(X_res2, y_res2, test_size=0.20, random_state=42)\n",
"\n",
"model_dt = DecisionTreeClassifier().fit(X_train_full_median, y_train_full_median)\n",
"probs2_dt = model_dt.predict_proba(X_test_full_median)#[:, 1]\n",
"\n",
"model_rf = RandomForestClassifier().fit(X_train_full_median, y_train_full_median)\n",
"probs2_rf = model_rf.predict_proba(X_test_full_median)#[:, 1]\n",
"\n",
"model_xg = XGBClassifier().fit(X_train_full_median, y_train_full_median)\n",
"probs2_xg = model_xg.predict_proba(X_test_full_median)#[:, 1]\n",
"\n",
"model_kn =KNeighborsClassifier().fit(X_train_full_median, y_train_full_median)\n",
"probs2_kn = model_xg.predict_proba(X_test_full_median)#[:, 1]\n",
"\n",
"model_SV = SVC(probability=True).fit(X_train_full_median, y_train_full_median)\n",
"probs2_SV = model_SV.predict_proba(X_test_full_median)\n",
"\n",
"\n",
"auc_dt2= roc_auc_score(y_test_full_median, probs2_dt, multi_class='ovr')\n",
"auc_rf2 = roc_auc_score(y_test_full_median, probs2_rf, multi_class='ovr')\n",
"auc_xg2 = roc_auc_score(y_test_full_median, probs2_xg, multi_class='ovr')\n",
"auc_kn2 = roc_auc_score(y_test_full_median, probs2_kn, multi_class='ovr')\n",
"auc_SV2 = roc_auc_score(y_test_full_median, probs2_SV, multi_class='ovr')\n",
"\n",
"print('- ROC_AUC of DTC: {:.4f}'.format(auc_dt2)) \n",
"print('- ROC_AUC of RFC: {:.4f}'.format(auc_rf2))\n",
"print('- ROC_AUC of XGB: {:.4f}'.format(auc_xg2))\n",
"print('- ROC_AUC of KNN: {:.4f}'.format(auc_kn2))\n",
"print('- ROC_AUC of SVM: {:.4f}'.format(auc_SV2))"
],
"metadata": {
"id": "Iv1P2HF6vagl",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "c4ea309b-42be-44b5-9b66-02f658843794"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"- ROC_AUC of DTC: 0.8929\n",
"- ROC_AUC of RFC: 0.9862\n",
"- ROC_AUC of XGB: 0.9751\n",
"- ROC_AUC of KNN: 0.9751\n",
"- ROC_AUC of SVM: 0.9514\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"\n",
"DTC = DecisionTreeClassifier()\n",
"RFC = RandomForestClassifier()\n",
"XGB = XGBClassifier()\n",
"KNN = KNeighborsClassifier()\n",
"SVM = SVC(probability=True)\n",
"\n",
"median2_cv_dt= cross_val_score(DTC, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"median2_cv_rf= cross_val_score(RFC, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"median2_cv_xg= cross_val_score(XGB, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"median2_cv_kn= cross_val_score(KNN, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"median2_cv_SV= cross_val_score(SVM, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').mean()\n",
"\n",
"median2_std_dt= cross_val_score(DTC, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"median2_std_rf= cross_val_score(RFC, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"median2_std_xg= cross_val_score(XGB, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"median2_std_kn= cross_val_score(KNN, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"median2_std_SV= cross_val_score(SVM, X_median_full, y_full, cv=10, scoring='roc_auc_ovr').std()\n",
"\n",
" \n",
"print('10-fold CV forDTC', (median2_cv_dt)) \n",
"print('10-fold CV forRFC', (median2_cv_rf))\n",
"print('10-fold CV forXGB',(median2_cv_xg))\n",
"print('10-fold CV forKNN',(median2_cv_kn))\n",
"print('10-fold CV forSVM',(median2_cv_SV))\n",
"\n",
"print('std forDTC', (median2_std_dt)) \n",
"print('std for RFC', (median2_std_rf))\n",
"print('std for XGB',(median2_std_xg))\n",
"print('std for KNN',(median2_std_kn))\n",
"print('std for SVM',(median2_std_SV))"
],
"metadata": {
"id": "x2tr5QyFvbP1",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "33623cbd-8560-4548-b922-d931b442bc39"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"10-fold CV forDTC 0.7643985387226901\n",
"10-fold CV forRFC 0.9225299487824339\n",
"10-fold CV forXGB 0.905462515029783\n",
"10-fold CV forKNN 0.8233527212676955\n",
"10-fold CV forSVM 0.8672833044572993\n",
"std forDTC 0.04192060406958924\n",
"std for RFC 0.027988450085827486\n",
"std for XGB 0.03213222959377166\n",
"std for KNN 0.05565549213845232\n",
"std for SVM 0.056710516142963456\n"
]
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "VgbT9-l4mkkC"
},
"source": [
"### Experiment: 3 = MICE for Imputing Null values"
]
},
{
"cell_type": "markdown",
"source": [],
"metadata": {
"id": "In5Oel1cneam"
}
},
{
"cell_type": "code",
"source": [
"full_data= df_final.copy()\n",
"X_full= full_data.drop('Phases', axis=1)\n",
"y_full = full_data['Phases']"
],
"metadata": {
"id": "N60dXbTDUZdi"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "9Ci2VDdgmkkC"
},
"outputs": [],
"source": [
"import miceforest as mf\n",
"\n",
"X_mice = X.copy()\n",
"kernel = mf.ImputationKernel(\n",
" X_mice,\n",
" save_all_iterations=True,\n",
" random_state=1989\n",
")# Run the MICE algorithm for 3 iterations kernel.mice(3)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "Zs81EaTXmkkC"
},
"outputs": [],
"source": [
"X_mice = kernel.complete_data()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "0aKXmwOumkkC"
},
"outputs": [],
"source": [
"# fit robust scaler\n",
"mice_pipeline = Pipeline(steps=[\n",
" ('RobustScaler', RobustScaler())\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "u9pKQrPmmkkC"
},
"outputs": [],
"source": [
"# Fit X with Mice imputer \n",
"X_mice= mice_pipeline.fit_transform(X_mice)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "dFb-8NDvmkkD"
},
"outputs": [],
"source": [
"# Resampling the minority class. The strategy can be changed as required.\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority', n_jobs=-1 )\n",
"# Fit the model to generate the data.\n",
"X_res2, y_res2 = smt.fit_resample(X_mice, y)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "d1ytZXJRmkkD",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "250bc020-5b57-4c79-bfcf-8fb4b0aaf1a3"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Random Forest\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.9176\n",
"- F1 score: 0.9181\n",
"- Precision: 0.9197\n",
"- Recall: 0.9176\n",
"===================================\n",
"\n",
"\n",
"Decision Tree\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8530\n",
"- F1 score: 0.8531\n",
"- Precision: 0.8536\n",
"- Recall: 0.8530\n",
"===================================\n",
"\n",
"\n",
"K-Neighbors Classifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.8293\n",
"- F1 score: 0.8277\n",
"- Precision: 0.8282\n",
"- Recall: 0.8293\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8029\n",
"- F1 score: 0.8000\n",
"- Precision: 0.8033\n",
"- Recall: 0.8029\n",
"===================================\n",
"\n",
"\n",
"XGBClassifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.9443\n",
"- F1 score: 0.9440\n",
"- Precision: 0.9442\n",
"- Recall: 0.9443\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8925\n",
"- F1 score: 0.8930\n",
"- Precision: 0.8971\n",
"- Recall: 0.8925\n",
"===================================\n",
"\n",
"\n",
"SVM\n",
"Model performance for Training set\n",
"- Accuracy: 0.8392\n",
"- F1 score: 0.8371\n",
"- Precision: 0.8421\n",
"- Recall: 0.8392\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8208\n",
"- F1 score: 0.8164\n",
"- Precision: 0.8293\n",
"- Recall: 0.8208\n",
"===================================\n",
"\n",
"\n"
]
}
],
"source": [
"# Training the models\n",
"report_mice = evaluate_models(X_res2, y_res2, models)"
]
},
{
"cell_type": "code",
"source": [
"mice_pipeline1 = Pipeline(steps=[\n",
" ('RobustScaler', RobustScaler())\n",
"])\n",
"\n",
"X_mice1= mice_pipeline1.fit_transform(X_full)\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority', n_jobs=-1 )\n",
"# Fit the model to generate the data.\n",
"X_res34, y_res34 = smt.fit_resample(X_mice1, y_full)\n",
"\n",
"X_train_mice12, X_test_mice12, y_train_mice12, y_test_mice12 = train_test_split(X_res34, y_res34, test_size=0.2, random_state=42)\n",
"\n",
"# model_rf_mice = RandomForestClassifier().fit(X_train_mice12, y_train_mice12)\n",
"\n",
"\n",
"model_dt = DecisionTreeClassifier().fit(X_train_mice12, y_train_mice12)\n",
"DT_mice_model_cb = model_dt.predict(X_test_mice12)\n",
"probs3_dt = model_dt.predict_proba(X_test_mice12)#[:, 1]\n",
"\n",
"model_rf = RandomForestClassifier().fit(X_train_mice12, y_train_mice12)\n",
"RF_mice_model_cb = model_rf.predict(X_test_mice12)\n",
"probs3_rf = model_rf.predict_proba(X_test_mice12)#[:, 1]\n",
"\n",
"model_xg = XGBClassifier().fit(X_train_mice12, y_train_mice12)\n",
"XG_mice_model_cb = model_xg.predict(X_test_mice12)\n",
"probs3_xg = model_xg.predict_proba(X_test_mice12)#[:, 1]\n",
"\n",
"model_kn =KNeighborsClassifier().fit(X_train_mice12, y_train_mice12)\n",
"KN_mice_model_cb = model_kn.predict(X_test_mice12)\n",
"probs3_kn = model_kn.predict_proba(X_test_mice12)#[:, 1]\n",
"\n",
"model_SV = SVC(probability=True).fit(X_train_mice12, y_train_mice12)\n",
"SVM_mice_model_cb = model_SV.predict(X_test_mice12)\n",
"probs3_SV = model_SV.predict_proba(X_test_mice12)\n",
"\n",
"#\n",
"auc_dt3 = roc_auc_score(y_test_mice12, probs3_dt, multi_class='ovr')\n",
"auc_rf3 = roc_auc_score(y_test_mice12, probs3_rf, multi_class='ovr')\n",
"auc_xg3 = roc_auc_score(y_test_mice12, probs3_xg , multi_class='ovr')\n",
"auc_kn3 = roc_auc_score(y_test_mice12, probs3_kn, multi_class='ovr')\n",
"auc_SV3 = roc_auc_score(y_test_mice12, probs3_SV, multi_class='ovr')\n",
"\n",
"#\n",
"print('- AUC of DTC: {:.4f}'.format(auc_dt3)) \n",
"print('- AUC of RFC: {:.4f}'.format(auc_rf3))\n",
"print('- AUC of XGB: {:.4f}'.format(auc_xg3))\n",
"print('- AUC of KNN: {:.4f}'.format(auc_kn3))\n",
"print('- AUC of SVM: {:.4f}'.format(auc_SV3))"
],
"metadata": {
"id": "BvPlGR--UFpo",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "98d0b131-9c90-4a52-c3e9-52e720c41bc4"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"- AUC of DTC: 0.8932\n",
"- AUC of RFC: 0.9861\n",
"- AUC of XGB: 0.9751\n",
"- AUC of KNN: 0.9282\n",
"- AUC of SVM: 0.9509\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"# sns.set_style('white')\n",
"# plt.figure(figsize=(9,8))\n",
"\n",
"# fpr1, tpr1, thresh1 = roc_curve(y_test_mice12, probs3_kn[:,1], pos_label=1)\n",
"# fpr2, tpr2, thresh2 = roc_curve(y_test_mice12, probs3_SV[:,1], pos_label=1)\n",
"# fpr3, tpr3, thresh3 = roc_curve(y_test_mice12, probs3_dt[:,1], pos_label=1) #DTC\n",
"# fpr4, tpr4, thresh4 = roc_curve(y_test_mice12, probs3_rf[:,1], pos_label=1)#RFC\n",
"# fpr6, tpr6, thresh6 = roc_curve(y_test_mice12, probs3_xg[:,1], pos_label=1) #XGB\n",
"\n",
"# a= roc_auc_score(y_test_mice12, probs3_kn, multi_class= 'ovr')\n",
"# b= roc_auc_score(y_test_mice12, probs3_SV, multi_class= 'ovr')\n",
"# c= roc_auc_score(y_test_mice12, probs3_dt, multi_class= 'ovr')\n",
"# d= roc_auc_score(y_test_mice12,probs3_rf, multi_class= 'ovr')\n",
"# e= roc_auc_score(y_test_mice12,probs3_xg, multi_class= 'ovr')\n",
"\n",
"# random_probs = [0 for i in range(len(y_test_mice12))]\n",
"# p_fpr, p_tpr, _ = roc_curve(y_test_mice12, random_probs, pos_label=1)\n",
"\n",
"# knn_roc_auc = a\n",
"# svm_roc_auc = b\n",
"# DT_roc_auc = c #auc(fpr3, tpr3)\n",
"# RF_roc_auc = d #auc(fpr4, tpr4)\n",
"# #GB_roc_auc = auc(fpr5, tpr5)\n",
"# XGB_roc_auc = e #auc(fpr6, tpr6)\n",
"\n",
"# plt.plot(fpr1, tpr1, linestyle= '--', color= 'red', label= 'ROC-AUC of ST-KNN = %0.4f' % knn_roc_auc, markersize=1)\n",
"# plt.plot(fpr2, tpr2 , linestyle= '--', color= 'green', label= 'ROC-AUC of ST-SVM = %0.4f' % svm_roc_auc, markersize=1)\n",
"# plt.plot(fpr3, tpr3, linestyle= '--', color= 'blue',label= 'ROC-AUC of ST-DTC =%0.4f' % DT_roc_auc, markersize=1 )\n",
"# plt.plot(fpr4, tpr4, linestyle= '--', color= 'purple', label= 'ROC-AUC of ST-RFC = %0.4f' % RF_roc_auc,markersize=1)\n",
"\n",
"# #plt.plot(fpr5, tpr5, linestyle= '--', color= 'cyan',label= 'AUC of Gradient Boost =%0.2f' % GB_roc_auc )\n",
"# plt.plot(fpr6, tpr6, linestyle= '--', color= 'magenta', label= 'ROC-AUC of ST-XGB = %0.4f' % XGB_roc_auc, markersize=1)\n",
"# plt.plot(p_fpr, p_tpr, linestyle= '--', color= 'black' )\n",
"# plt.title(' ROC Curve for all five SMOTE-Tomek models', fontsize=20)\n",
"# plt.xlabel('False positive rate', fontsize=20)\n",
"# plt.ylabel('True positive rate', fontsize=20)\n",
"# plt.legend(loc='best', fontsize=15)\n",
"# plt.xticks( fontsize=15)\n",
"# plt.yticks(fontsize=15)\n",
"# #plt.savefig('ROC', dpi=300)"
],
"metadata": {
"id": "goRz0DsEYq2f"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"metadata": {
"id": "mHPzp3N6mkkD"
},
"source": [
"### Experiment: 4 = Simple Imputer with Strategy Constant "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "_6TzxxAomkkD"
},
"outputs": [],
"source": [
"# Create a pipeline with simple imputer with strategy constant and fill value 0\n",
"constant_pipeline = Pipeline(steps=[\n",
" ('Imputer', SimpleImputer(strategy='constant', fill_value=0)),\n",
" ('RobustScaler', RobustScaler())\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "vztriY6kmkkD"
},
"outputs": [],
"source": [
"X_const =constant_pipeline.fit_transform(X)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "Gn4rGwRwmkkD"
},
"outputs": [],
"source": [
"# Resampling the minority class. The strategy can be changed as required.\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority', n_jobs=-1 )\n",
"# Fit the model to generate the data.\n",
"X_res3, y_res3 = smt.fit_resample(X_const, y)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "eGPMRNjjmkkE",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "a5af99da-c4df-46d3-a762-f4683f400ff1"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Random Forest\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.9068\n",
"- F1 score: 0.9070\n",
"- Precision: 0.9081\n",
"- Recall: 0.9068\n",
"===================================\n",
"\n",
"\n",
"Decision Tree\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8459\n",
"- F1 score: 0.8465\n",
"- Precision: 0.8483\n",
"- Recall: 0.8459\n",
"===================================\n",
"\n",
"\n",
"K-Neighbors Classifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.8293\n",
"- F1 score: 0.8277\n",
"- Precision: 0.8282\n",
"- Recall: 0.8293\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8029\n",
"- F1 score: 0.8000\n",
"- Precision: 0.8033\n",
"- Recall: 0.8029\n",
"===================================\n",
"\n",
"\n",
"XGBClassifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.9443\n",
"- F1 score: 0.9440\n",
"- Precision: 0.9442\n",
"- Recall: 0.9443\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8925\n",
"- F1 score: 0.8930\n",
"- Precision: 0.8971\n",
"- Recall: 0.8925\n",
"===================================\n",
"\n",
"\n",
"SVM\n",
"Model performance for Training set\n",
"- Accuracy: 0.8392\n",
"- F1 score: 0.8371\n",
"- Precision: 0.8421\n",
"- Recall: 0.8392\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8208\n",
"- F1 score: 0.8164\n",
"- Precision: 0.8293\n",
"- Recall: 0.8208\n",
"===================================\n",
"\n",
"\n"
]
}
],
"source": [
"# training the models\n",
"report_const = evaluate_models(X_res3, y_res3, models)"
]
},
{
"cell_type": "code",
"source": [
"full_data= df_final.copy()\n",
"X_full= full_data.drop('Phases', axis=1)\n",
"y_full = full_data['Phases']"
],
"metadata": {
"id": "CqgUrDjButSW"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"constant_pipeline1 = Pipeline(steps=[\n",
" ('Imputer', SimpleImputer(strategy='constant', fill_value=0)),\n",
" ('RobustScaler', RobustScaler())])\n",
"\n",
"X_constan1= constant_pipeline1.fit_transform(X_full)\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority', n_jobs=-1 )\n",
"# Fit the model to generate the data.\n",
"X_res34, y_res34 = smt.fit_resample(X_constan1, y_full)\n",
"\n",
"X_train_constan12, X_test_constan12, y_train_constan12, y_test_constan12 = train_test_split(X_res34, y_res34, test_size=0.2, random_state=42)\n",
"\n",
"model_rf_constan = RandomForestClassifier().fit(X_train_constan12, y_train_constan12)"
],
"metadata": {
"id": "0Od8eSFXujBO"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"X_res34.shape"
],
"metadata": {
"id": "bSo2-ab8vfmw",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "55a782d7-cdc4-4a70-9ff1-ca9108407233"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(1392, 35)"
]
},
"metadata": {},
"execution_count": 952
}
]
},
{
"cell_type": "code",
"source": [
"y_res34.value_counts()"
],
"metadata": {
"id": "FNdRriv4vjHJ",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "1f8ccd28-d625-4b3f-806b-9598df804fd5"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"3.0 434\n",
"0.0 408\n",
"1.0 350\n",
"2.0 200\n",
"Name: Phases, dtype: int64"
]
},
"metadata": {},
"execution_count": 953
}
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "00cLvBTlmkkE"
},
"source": [
"## Experiment: 5 = Simple Imputer with Strategy Mean \n",
"\n",
"- Another strategy which can be used is replacing missing values with mean\n",
"- Here we replace the missing values with the mean of the column"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "kPxglohumkkE"
},
"outputs": [],
"source": [
"# Create a pipeline with Simple imputer with strategy mean\n",
"mean_pipeline = Pipeline(steps=[\n",
" ('Imputer', SimpleImputer(strategy='mean')),\n",
" ('RobustScaler', RobustScaler())\n",
"])"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "f1Kv0g93mkkE"
},
"outputs": [],
"source": [
"X_mean = mean_pipeline.fit_transform(X)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "zj6nWgwrmkkE"
},
"outputs": [],
"source": [
"# Resampling the minority class. The strategy can be changed as required.\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority' , n_jobs=-1)\n",
"# Fit the model to generate the data.\n",
"X_res4, y_res4 = smt.fit_resample(X_mean, y)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "OiW9jYe7mkkE",
"colab": {
"base_uri": "https://localhost:8080/"
},
"outputId": "d04b1d28-c5dd-4275-d4c5-4d51138c6602"
},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Random Forest\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.9140\n",
"- F1 score: 0.9138\n",
"- Precision: 0.9159\n",
"- Recall: 0.9140\n",
"===================================\n",
"\n",
"\n",
"Decision Tree\n",
"Model performance for Training set\n",
"- Accuracy: 1.0000\n",
"- F1 score: 1.0000\n",
"- Precision: 1.0000\n",
"- Recall: 1.0000\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8459\n",
"- F1 score: 0.8467\n",
"- Precision: 0.8493\n",
"- Recall: 0.8459\n",
"===================================\n",
"\n",
"\n",
"K-Neighbors Classifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.8293\n",
"- F1 score: 0.8277\n",
"- Precision: 0.8282\n",
"- Recall: 0.8293\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8029\n",
"- F1 score: 0.8000\n",
"- Precision: 0.8033\n",
"- Recall: 0.8029\n",
"===================================\n",
"\n",
"\n",
"XGBClassifier\n",
"Model performance for Training set\n",
"- Accuracy: 0.9443\n",
"- F1 score: 0.9440\n",
"- Precision: 0.9442\n",
"- Recall: 0.9443\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8925\n",
"- F1 score: 0.8930\n",
"- Precision: 0.8971\n",
"- Recall: 0.8925\n",
"===================================\n",
"\n",
"\n",
"SVM\n",
"Model performance for Training set\n",
"- Accuracy: 0.8392\n",
"- F1 score: 0.8371\n",
"- Precision: 0.8421\n",
"- Recall: 0.8392\n",
"----------------------------------\n",
"Model performance for Test set\n",
"- Accuracy: 0.8208\n",
"- F1 score: 0.8164\n",
"- Precision: 0.8293\n",
"- Recall: 0.8208\n",
"===================================\n",
"\n",
"\n"
]
}
],
"source": [
"# Training all models\n",
"report_mean = evaluate_models(X_res4, y_res4, models)"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "hCflOLkYmkkG"
},
"outputs": [],
"source": [
"mean_pipeline1 = Pipeline(steps=[\n",
" ('Imputer', SimpleImputer(strategy='mean')),\n",
" ('RobustScaler', RobustScaler())\n",
"])\n",
"\n",
"\n",
"X_mean1= mean_pipeline1.fit_transform(X_full)\n",
"smt = SMOTETomek(random_state=42,sampling_strategy='minority', n_jobs=-1 )\n",
"# Fit the model to generate the data.\n",
"X_res35, y_res35 = smt.fit_resample(X_mean1, y_full)\n",
"\n",
"X_train_mean12, X_test_mean12, y_train_mean12, y_test_mean12 = train_test_split(X_res35, y_res35, test_size=0.2, random_state=42)\n",
"\n",
"model_rf_mean = RandomForestClassifier().fit(X_train_mean12, y_train_mean12)\n",
"RF_Prob = model_rf_mean.predict_proba(X_test_mean12)\n",
"\n",
"# from sklearn.metrics import plot_confusion_matrix\n",
"# class_names = ['IM', 'BCC', 'FCC', 'FCC+BCC']\n",
"# cmplot= plot_confusion_matrix(model_rf_mean, X_test_mean12, y_test_mean12, display_labels= class_names, cmap='gist_heat')\n",
"# cmplot.ax_.set_title('Confusion Matrix', color='red')\n",
"# plt.xlabel('Predicted Label', color='black')\n",
"# plt.ylabel('True Label', color='black')\n",
"# plt.gcf().axes[0].tick_params(color='black')\n",
"# plt.gcf().axes[1].tick_params(color='black')\n",
"# plt.gcf().set_size_inches(10,6)\n",
"# plt.show()"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {
"id": "MI-BqUpNmkkG"
},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3 (ipykernel)",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.9.12"
},
"vscode": {
"interpreter": {
"hash": "b7082a90f0341f66b325168da8fc238f0b2aba7ee16848d917086bb4ed45c134"
}
},
"colab": {
"provenance": []
}
},
"nbformat": 4,
"nbformat_minor": 0
}