Merge pull request #1693 from Gokul45-45/patch-4

feat: Add evaluation metrics module with accuracy, precision, and rec…

Author: sanjay-kv

projects/detection/Medical_Appointment_No_Shows/evaluate_metrics.py

@@ -0,0 +1,99 @@
+"""Evaluation metrics module for Medical Appointment No-Shows model.
+
+This module provides comprehensive evaluation metrics including accuracy,
+precision, recall, F1-score, confusion matrix, and ROC-AUC score.
+"""
+
+import numpy as np
+import matplotlib.pyplot as plt
+from sklearn.metrics import (
+    accuracy_score, precision_score, recall_score, f1_score,
+    confusion_matrix, roc_auc_score, classification_report,
+    roc_curve, auc
+)
+import seaborn as sns
+
+
+class MetricsEvaluator:
+    """Evaluate machine learning model performance with comprehensive metrics."""
+    
+    def __init__(self):
+        self.metrics = {}
+        self.predictions = None
+        self.actual = None
+    
+    def calculate_metrics(self, y_true, y_pred, y_pred_proba=None):
+        """Calculate comprehensive evaluation metrics.
+        
+        Args:
+            y_true: True labels
+            y_pred: Predicted labels
+            y_pred_proba: Predicted probabilities (for ROC-AUC)
+            
+        Returns:
+            dict: Dictionary containing all evaluation metrics
+        """
+        self.actual = y_true
+        self.predictions = y_pred
+        
+        # Calculate basic metrics
+        self.metrics['accuracy'] = accuracy_score(y_true, y_pred)
+        self.metrics['precision'] = precision_score(y_true, y_pred, average='weighted', zero_division=0)
+        self.metrics['recall'] = recall_score(y_true, y_pred, average='weighted', zero_division=0)
+        self.metrics['f1_score'] = f1_score(y_true, y_pred, average='weighted', zero_division=0)
+        
+        # Calculate ROC-AUC if probabilities are provided
+        if y_pred_proba is not None:
+            try:
+                self.metrics['roc_auc'] = roc_auc_score(y_true, y_pred_proba)
+            except:
+                self.metrics['roc_auc'] = None
+        
+        # Confusion matrix
+        self.metrics['confusion_matrix'] = confusion_matrix(y_true, y_pred)
+        
+        return self.metrics
+    
+    def print_report(self):
+        """Print detailed classification report."""
+        if self.actual is None or self.predictions is None:
+            print("No metrics calculated yet. Call calculate_metrics first.")
+            return
+        
+        print("\n=== Classification Report ===")
+        print(classification_report(self.actual, self.predictions))
+    
+    def display_metrics(self):
+        """Display all calculated metrics."""
+        print("\n=== Evaluation Metrics ===")
+        print(f"Accuracy:  {self.metrics.get('accuracy', 'N/A'):.4f}")
+        print(f"Precision: {self.metrics.get('precision', 'N/A'):.4f}")
+        print(f"Recall:    {self.metrics.get('recall', 'N/A'):.4f}")
+        print(f"F1-Score:  {self.metrics.get('f1_score', 'N/A'):.4f}")
+        if self.metrics.get('roc_auc'):
+            print(f"ROC-AUC:   {self.metrics.get('roc_auc', 'N/A'):.4f}")
+    
+    def plot_confusion_matrix(self, save_path=None):
+        """Plot confusion matrix heatmap.
+        
+        Args:
+            save_path: Path to save the plot (optional)
+        """
+        cm = self.metrics.get('confusion_matrix')
+        if cm is None:
+            print("No confusion matrix available.")
+            return
+        
+        plt.figure(figsize=(8, 6))
+        sns.heatmap(cm, annot=True, fmt='d', cmap='Blues')
+        plt.title('Confusion Matrix')
+        plt.ylabel('True Label')
+        plt.xlabel('Predicted Label')
+        
+        if save_path:
+            plt.savefig(save_path)
+        plt.show()
+
+
+if __name__ == '__main__':
+    print('Metrics Evaluator module for Medical Appointment No-Shows prediction model')