Merge pull request #116 from darshbaxi/darsh/rmse

Add Root Mean Squared Error (RMSE) metric to pydeepflow.metrics  (#111)

Author: ravin-d-27

pydeepflow/metrics.py

@@ -157,12 +157,12 @@ def r2_score(y_true, y_pred):
     ss_tot = np.sum((y_true - np.mean(y_true)) ** 2)
     return 1 - (ss_res / ss_tot)
 
-def rmse(y_true, y_pred):
+def root_mean_squared_error(y_true, y_pred):
     """
-    Compute Root Mean Squared Error between true and predicted values.
-    
+    Calculates the Root Mean Squared Error (RMSE).
+
     RMSE = sqrt((1/n) * Σ(y_true - y_pred)^2)
-    
+
     Parameters
     ----------
     y_true : array-like
@@ -175,7 +175,5 @@ def rmse(y_true, y_pred):
     float
         The RMSE score.
     """
-    y_true = np.array(y_true)
-    y_pred = np.array(y_pred)
-    return np.sqrt(np.mean((y_true - y_pred) ** 2))
-
+    return ((np.array(y_true) - np.array(y_pred)) ** 2).mean() ** 0.5
+    

pydeepflow/model.py

@@ -4,7 +4,7 @@
 import matplotlib.pyplot as plt
 from pydeepflow.activations import activation, activation_derivative
 from pydeepflow.losses import get_loss_function, get_loss_derivative
-from pydeepflow.metrics import precision_score, recall_score, f1_score, confusion_matrix,mean_absolute_error, mean_squared_error, r2_score
+from pydeepflow.metrics import precision_score, recall_score, f1_score, confusion_matrix,mean_absolute_error, mean_squared_error, r2_score, root_mean_squared_error
 from pydeepflow.device import Device
 from pydeepflow.regularization import Regularization
 from pydeepflow.checkpoints import ModelCheckpoint
@@ -817,7 +817,7 @@ def evaluate(self, X, y, metrics=['loss', 'accuracy']):
             y (np.ndarray): The true labels for evaluation.
             metrics (list, optional): A list of metrics to calculate.
                                       Defaults to ['loss', 'accuracy'].
-                                      Available metrics: 'loss', 'accuracy', 'precision', 'recall', 'f1_score', 'confusion_matrix'.
+                                      Available metrics: 'loss', 'accuracy', 'precision', 'recall', 'f1_score', 'confusion_matrix', 'root_mean_squared_error'.
 
         Returns:
             dict: A dictionary where keys are the metric names and values are the computed scores.
@@ -856,6 +856,9 @@ def evaluate(self, X, y, metrics=['loss', 'accuracy']):
         if 'r2_score' in metrics:
          results['r2_score'] = r2_score(y, predictions)
 
+        if 'root_mean_squared_error' in metrics:
+            results['root_mean_squared_error'] = root_mean_squared_error(y, predictions)
+
 
 
         return results
@@ -1740,6 +1743,9 @@ def evaluate(self, X, y, metrics=['loss', 'accuracy']):
 
         if 'r2_score' in metrics:
             results['r2_score'] = r2_score(y, predictions)    
+        
+        if 'root_mean_squared_error' in metrics:
+            results['root_mean_squared_error'] = root_mean_squared_error(y, predictions)
 
         return results # Removed confusion_matrix for simplification/dependency reasons
 

tests/test_metrics.py

@@ -2,7 +2,8 @@
 import numpy as np
 from pydeepflow.metrics import (
     precision_score, recall_score, f1_score, confusion_matrix,
-    mean_absolute_error, mean_squared_error, r2_score, rmse
+    mean_absolute_error, mean_squared_error, r2_score,root_mean_squared_error  
+
 )
 
 class TestMetrics(unittest.TestCase):
@@ -54,11 +55,23 @@ def test_r2_score(self):
         # R^2 = 1 - (1.5 / 29.1875) = 1 - 0.051389... approx 0.9486
         self.assertAlmostEqual(r2_score(self.y_true_reg, self.y_pred_reg), 0.94861051, places=5)
 
-    def test_rmse(self):
-        y_true = np.array([1, 2, 3])
-        y_pred = np.array([2, 2, 4])
-        expected = np.sqrt(((1-2)**2 + (2-2)**2 + (3-4)**2) / 3)
-        self.assertAlmostEqual(rmse(y_true, y_pred), expected)
+
+    def test_root_mean_squared_error(self):
+        # Step 1: Differences
+        # (3 - 2.5) = 0.5
+        # (-0.5 - 0.0) = -0.5
+        # (2 - 2) = 0
+        # (7 - 8) = -1
+        #
+        # Step 2: Squared differences
+        # [0.5², (-0.5)², 0², (-1)²] = [0.25, 0.25, 0, 1]
+        #
+        # Step 3: Mean Squared Error (MSE)
+        # (0.25 + 0.25 + 0 + 1) / 4 = 0.375
+        #
+        # Step 4: Root Mean Squared Error (RMSE)
+        # sqrt(0.375) = 0.6123724356957945
+        self.assertAlmostEqual(root_mean_squared_error(self.y_true_reg, self.y_pred_reg), 0.6123724356957945, places=6)
 
 if __name__ == '__main__':
     unittest.main()