Merge pull request #16 from rugbedbugg/issue_3_branch

fixed issue #3

neo.py

@@ -11,7 +11,21 @@ def sigmoid(z):
         return 1.0 if z > 0 else 0.0
 
 class LogisticRegression:
+issue_3_branch
+    def compute_loss(self, X, y):
+        linear_model = np.dot(X, self.weights) + self.bias
+        y_predicted = sigmoid(linear_model)
+        log_loss = -np.mean(y * np.log(y_predicted) + (1 - y) * np.log(1 - y_predicted))
+
+        if self.use_regularization:
+            log_loss += (self.regularization_strength / 2) * np.sum(np.square(self.weights))  # L2 regularization
+
+        return log_loss
+
+    def __init__(self, learning_rate=0.01, epochs=50, batch_size=4, regularization_strength=0.01, use_regularization=True):
+=======
     def __init__(self, learning_rate=0.01, epochs=50, batch_size=4, regularization_strength=0.01, use_regularization=True, learning_rate_deacy = 0.99):
+main
         self.learning_rate = learning_rate
         self.epochs = epochs
         self.batch_size = batch_size
@@ -84,6 +98,9 @@ def fit(self, X, y):
 =======
             self.learning_rate *= self.learning_rate_decay
 
+            loss = self.compute_loss(X, y)
+            print(f'Epoch {epoch+1}/{self.epochs}, Loss: {loss:.4f}')            
+
             if np.allclose(prev_weights, self.weights, rtol=1e-05):  # Corrected stopping condition
                 break