Create Python Script

Python Script

@@ -0,0 +1,157 @@
+# Complete Python Script: From EDA to Feature Engineering to Decision Tree Classifier
+
+# Step 1: Import necessary libraries
+import pandas as pd
+import numpy as np
+import matplotlib.pyplot as plt
+import seaborn as sns
+from sklearn.preprocessing import StandardScaler
+from sklearn.model_selection import train_test_split
+from sklearn.tree import DecisionTreeClassifier, plot_tree
+from sklearn.metrics import confusion_matrix, classification_report, accuracy_score
+
+# Step 2: Load the dataset
+file_path = '/mnt/data/Loan_Modelling (1).csv'  # Replace this with the correct file path in Colab
+data = pd.read_csv(file_path)
+
+# Step 3: Exploratory Data Analysis (EDA)
+
+# Check for missing values
+missing_values = data.isnull().sum()
+print("Missing Values:\n", missing_values)
+
+# Univariate Analysis
+def univariate_analysis(data, column, chart_type='histogram'):
+    if chart_type == 'histogram':
+        plt.figure(figsize=(8, 6))
+        sns.histplot(data[column], kde=True, bins=30)
+        plt.title(f"Distribution of {column}")
+        plt.show()
+    elif chart_type == 'boxplot':
+        plt.figure(figsize=(8, 6))
+        sns.boxplot(data[column])
+        plt.title(f"Boxplot of {column}")
+        plt.show()
+
+# Plot for numeric variables
+numerical_columns = ['Age', 'Experience', 'Income', 'CCAvg', 'Mortgage']
+for col in numerical_columns:
+    univariate_analysis(data, col)
+
+# Multivariate Analysis
+plt.figure(figsize=(10, 8))
+sns.heatmap(data.corr(), annot=True, fmt=".2f", cmap="coolwarm")
+plt.title("Correlation Heatmap")
+plt.show()
+
+# Scatter plots for visualizing relationships
+plt.figure(figsize=(10, 6))
+sns.scatterplot(data=data, x='Income', y='CCAvg', hue='Personal_Loan')
+plt.title('Income vs. Credit Card Spending by Loan Status')
+plt.show()
+
+plt.figure(figsize=(10, 6))
+sns.scatterplot(data=data, x='Income', y='Mortgage', hue='Personal_Loan')
+plt.title('Income vs. Mortgage by Loan Status')
+plt.show()
+
+# Step 4: Data Preprocessing and Feature Engineering
+
+# Drop unnecessary columns
+data = data.drop(['ID', 'ZIPCode'], axis=1)
+
+# Handle missing values
+data.fillna(data.median(), inplace=True)
+
+# Convert numeric columns to appropriate types
+numeric_columns = ['Age', 'Experience', 'Income', 'CCAvg', 'Mortgage']
+for col in numeric_columns:
+    data[col] = pd.to_numeric(data[col], errors='coerce')
+data.fillna(data.median(), inplace=True)  # Handle resulting NaNs
+
+# Feature Engineering
+data['Income_to_Mortgage_Ratio'] = data['Mortgage'] / (data['Income'] + 1e-5)
+data['Total_Accounts'] = data['Securities_Account'] + data['CD_Account'] + data['CreditCard']
+data['Spending_to_Income_Ratio'] = data['CCAvg'] / (data['Income'] + 1e-5)
+
+# Binning income into brackets
+income_bins = [data['Income'].min(), data['Income'].quantile(0.33), data['Income'].quantile(0.66), data['Income'].max()]
+income_labels = ['Low', 'Medium', 'High']
+data['Income_Bracket'] = pd.cut(data['Income'], bins=income_bins, labels=income_labels)
+
+# Binning age into groups
+age_bins = [data['Age'].min(), 35, 55, data['Age'].max()]
+age_labels = ['Young', 'Middle Aged', 'Senior']
+data['Age_Group'] = pd.cut(data['Age'], bins=age_bins, labels=age_labels)
+
+# Scale continuous features
+scaler = StandardScaler()
+scaled_features = ['Income', 'CCAvg', 'Mortgage', 'Income_to_Mortgage_Ratio', 'Spending_to_Income_Ratio']
+data[scaled_features] = scaler.fit_transform(data[scaled_features])
+
+# Encode categorical variables
+categorical_features = ['Education', 'Family', 'Income_Bracket', 'Age_Group']
+data = pd.get_dummies(data, columns=categorical_features, drop_first=True)
+
+# Step 5: Prepare Data for Modeling
+X = data.drop('Personal_Loan', axis=1)
+y = data['Personal_Loan']
+
+# Split into training and testing sets
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42, stratify=y)
+
+# Step 6: Decision Tree Classifier with Weighted Classes
+dt_classifier_weighted = DecisionTreeClassifier(
+    criterion='gini',
+    max_depth=None,
+    class_weight={0: 1, 1: 5},
+    random_state=42
+)
+dt_classifier_weighted.fit(X_train, y_train)
+
+# Step 7: Model Evaluation
+y_pred_weighted = dt_classifier_weighted.predict(X_test)
+
+# Confusion Matrix
+conf_matrix_weighted = confusion_matrix(y_test, y_pred_weighted)
+print("Confusion Matrix (Weighted):\n", conf_matrix_weighted)
+
+# Visualize Confusion Matrix
+plt.figure(figsize=(6, 4))
+sns.heatmap(conf_matrix_weighted, annot=True, fmt='d', cmap='Blues', xticklabels=['No Loan', 'Loan'], yticklabels=['No Loan', 'Loan'])
+plt.ylabel('Actual')
+plt.xlabel('Predicted')
+plt.title('Confusion Matrix (Weighted)')
+plt.show()
+
+# Classification Report
+class_report_weighted = classification_report(y_test, y_pred_weighted)
+print("Classification Report (Weighted):\n", class_report_weighted)
+
+# Accuracy Score
+accuracy_weighted = accuracy_score(y_test, y_pred_weighted)
+print("Accuracy of the Weighted Decision Tree model:", accuracy_weighted)
+
+# Feature Importance
+feature_importances_weighted = pd.Series(dt_classifier_weighted.feature_importances_, index=X.columns)
+feature_importances_weighted = feature_importances_weighted.sort_values(ascending=False)
+
+print("Feature Importances (Weighted):\n", feature_importances_weighted)
+
+# Visualize Feature Importances
+plt.figure(figsize=(12, 6))
+feature_importances_weighted.plot(kind='bar')
+plt.title('Feature Importances (Weighted)')
+plt.show()
+
+# Visualize the Decision Tree
+plt.figure(figsize=(20, 10))
+plot_tree(
+    dt_classifier_weighted,
+    feature_names=X.columns,
+    class_names=['No Loan', 'Loan'],
+    filled=True,
+    rounded=True,
+    fontsize=8
+)
+plt.show()