Evaluation Metrics for a Classification Approach with Decision Tree Algorithm

 #Complete Evaluation Metrics for a Classification Approach with Decision Tree Algorithm 

# Step 1: Import libraries and generate synthetic binary classification dataset from sklearn.datasets import make_classification from sklearn.model_selection import train_test_split # Generate a dataset with 1000 samples, 2 classes (binary), with slight imbalance X, y = make_classification( n_samples=1000, n_classes=2, weights=[0.6, 0.4], # 60% class 0, 40% class 1 n_features=2, n_redundant=0, n_clusters_per_class=1, random_state=42 ) # Initial split: 80% for training+validation, 20% for test X_temp, X_test, y_temp, y_test = train_test_split(X, y, test_size=0.2, random_state=42) # From the 80%, split into 60% training and 20% validation (i.e., 75% train / 25% val of X_temp) X_train, X_val, y_train, y_val = train_test_split(X_temp, y_temp, test_size=0.25, random_state=42) # Check sizes print(f"Training set: {X_train.shape}, Validation set: {X_val.shape}, Test set: {X_test.shape}")

#Output: Training set: (600, 2), Validation set: (200, 2), Test set: (200, 2)

from sklearn.tree import DecisionTreeClassifier # Initialize and train Decision Tree clf = DecisionTreeClassifier(random_state=42) clf.fit(X_train, y_train) # Predict on validation set y_pred_val = clf.predict(X_val) y_prob_val = clf.predict_proba(X_val)[:, 1] # probabilities for ROC curve # Show predictions shape for confirmation print("Predictions shape:", y_pred_val.shape) #Output: Predictions shape: (200,)

#Counfusion Matrix import matplotlib.pyplot as plt import seaborn as sns import numpy as np from sklearn.metrics import confusion_matrix # Compute confusion matrix cm = confusion_matrix(y_val, y_pred_val) # Title for the confusion matrix title = "Confusion Matrix - DValidation Set" # Set font properties globally plt.rcParams.update({ 'font.size': 18, 'font.family': 'serif', 'axes.titlesize': 18, 'axes.labelsize': 18, 'xtick.labelsize': 18, 'ytick.labelsize': 18 }) # Create figure fig, ax = plt.subplots(figsize=(8, 4)) # Define color map cmap = sns.color_palette("crest", as_cmap=True) # Plot heatmap with borders sns.heatmap(cm, annot=True, fmt='d', cmap=cmap, cbar=True, ax=ax, annot_kws={"fontsize": 18}, linewidths=0.5, linecolor='white') # Axis labels and ticks ax.set_title(title) ax.set_xlabel("Predicted Labels") ax.set_ylabel("True Labels") ax.set_xticklabels(["Class 0", "Class 1"], rotation=45, fontsize=14) ax.set_yticklabels(["Class 0", "Class 1"], rotation=0, fontsize=14) # Gridlines inside the heatmap ax.hlines([1], *ax.get_xlim(), colors='white', linewidth=4) ax.vlines([1], *ax.get_ylim(), colors='white', linewidth=4) # Show plot plt.tight_layout() plt.show()

#Other Metrics from sklearn.metrics import ( confusion_matrix, accuracy_score, precision_score, recall_score, f1_score, roc_auc_score, balanced_accuracy_score, matthews_corrcoef ) # Compute confusion matrix cm = confusion_matrix(y_val, y_pred_val) TN, FP, FN, TP = cm.ravel() # Core metrics accuracy = accuracy_score(y_val, y_pred_val) precision = precision_score(y_val, y_pred_val) recall = recall_score(y_val, y_pred_val) # TPR f1 = f1_score(y_val, y_pred_val) roc_auc = roc_auc_score(y_val, y_prob_val) # Additional metrics specificity = TN / (TN + FP) # TNR fpr = FP / (FP + TN) # FPR fnr = FN / (FN + TP) # FNR balanced_acc = balanced_accuracy_score(y_val, y_pred_val) mcc = matthews_corrcoef(y_val, y_pred_val) # Print all metrics print(f"\nAccuracy: {accuracy:.4f}") print(f"Precision (PPV): {precision:.4f}") print(f"Recall (Sensitivity): {recall:.4f}") print(f"F1 Score: {f1:.4f}") print(f"ROC AUC: {roc_auc:.4f}") print(f"Specificity (TNR): {specificity:.4f}") print(f"False Positive Rate: {fpr:.4f}") print(f"False Negative Rate: {fnr:.4f}") print(f"Balanced Accuracy: {balanced_acc:.4f}") print(f"Matthews Corr Coef: {mcc:.4f}")

from sklearn.metrics import classification_report, roc_curve import matplotlib.pyplot as plt # Classification report report = classification_report(y_val, y_pred_val, target_names=["Class 0", "Class 1"]) print("Classification Report:\n", report)

# ROC Curve fpr, tpr, thresholds = roc_curve(y_val, y_prob_val) # Plot plt.figure(figsize=(6, 4)) plt.plot(fpr, tpr, label=f'ROC Curve (AUC = {roc_auc:.2f})', color='blue') plt.plot([0, 1], [0, 1], linestyle='--', color='gray', label='Random Guessing') plt.xlabel('False Positive Rate') plt.ylabel('True Positive Rate (Recall)') plt.title('ROC Curve - Decision Tree (Validation Set)') plt.legend() plt.grid(True) plt.tight_layout() plt.show()

#Cross Validation from sklearn.model_selection import StratifiedKFold, cross_val_predict from sklearn.metrics import roc_auc_score # Define 5-fold stratified cross-validation cv = StratifiedKFold(n_splits=5, shuffle=True, random_state=42) # Predict probabilities using cross-validation y_cv_prob = cross_val_predict(clf, X_train, y_train, cv=cv, method='predict_proba')[:, 1] # ROC AUC from cross-validated predictions cv_auc = roc_auc_score(y_train, y_cv_prob) print(f"Cross-Validation ROC AUC (5-fold): {cv_auc:.4f}")

#Output: Cross-Validation ROC AUC (5-fold): 0.9216

#Thank you: Good Luck!!!