更新 function.py

function.py

@@ -29,6 +29,15 @@ name = "Tanjunfeng"
 # 这里写自己的代码实现
 
 def sigmoid(z):
+    """
+    Compute the sigmoid of z.
+
+    Args:
+    - z (numpy array): Input array.
+
+    Returns:
+    - numpy array: Sigmoid of input.
+    """
     return 1 / (1 + np.exp(-z))
 
 def train_logreg(X, y, learning_rate=0.01, iterations=1000):
@@ -42,7 +51,7 @@ def train_logreg(X, y, learning_rate=0.01, iterations=1000):
     - iterations (int): 迭代次数，默认为 1000。
 
     Returns:
-    - numpy array: 训练得到的权重。
+    - Tuple[numpy array, float]: 训练得到的权重和偏置。
     """
     m, n = X.shape
     weights = np.zeros(n)
@@ -56,23 +65,23 @@ def train_logreg(X, y, learning_rate=0.01, iterations=1000):
         weights -= learning_rate * dw
         bias -= learning_rate * db
 
-    return weights
+    return weights, bias
 
-def predict_logreg(inX, weights):
+def predict_logreg(inX, weights, bias):
     """
     对率回归预测函数。
 
     Args:
     - inX (numpy array): 输入待分类的样本。
     - weights (numpy array): 训练得到的权重。
+    - bias (float): 训练得到的偏置。
 
     Returns:
     - int: 预测的分类标签。
     """
-    z = np.dot(inX, weights)
+    z = np.dot(inX, weights) + bias
     h = sigmoid(z)
-    # 调整分类阈值
-    return 1 if h > 0.6 else 0
+    return 1 if h >= 0.5 else 0
 
 def load_data(filepath):
     """
@@ -82,23 +91,22 @@ def load_data(filepath):
     - directory (str): 数据文件路径。
 
     Returns:
-    - np.array: 特征数据。
-    - np.array: 标签数据。
+    - Tuple[np.array, np.array]: 特征数据和标签数据。
     """
     data = []
     with open(filepath, 'r') as f:
         for line in f:
             row = []
             values = line.strip().split()
+            if '?' in values:
+                continue  # Ignore lines with missing data
             for value in values:
-                try:
-                    row.append(float(value))
-                except ValueError:
-                    row.append(-1)
+                row.append(float(value))
             data.append(row)
     np_data = np.array(data)
     X = np_data[:, :-1]
     y = np_data[:, -1]
+    y = np.array([0 if label == 1 else (1 if label == 2 else -1) for label in y])
     return X, y
 
 if __name__ == "__main__":
@@ -110,14 +118,14 @@ if __name__ == "__main__":
     X_test, y_test = load_data("./data/horse-colic.test")
 
     # 训练对率回归模型
-    weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=5000)
+    weights, bias = train_logreg(X_train, y_train, learning_rate=0.001, iterations=5000)
 
     # 测试单个数据
     correct_count = 0
     total_count = len(y_test)
     for i in range(total_count):
-        y_pred = predict_logreg(X_test[i], weights)
+        y_pred = predict_logreg(X_test[i], weights, bias)
         if y_pred == y_test[i]:
             correct_count += 1
     accuracy = correct_count / total_count
-    print(f"准确率: {accuracy}")
+    print(f"测试准确率: {accuracy:.2f}")