更新 function.py

function.py

@@ -25,32 +25,39 @@ def sigmoid(z):
 """
 id = "202210423223"
 name = "李顺潮"
+
 # 对率回归训练函数  
-def train_logreg(X, y, learning_rate=0.01, iterations=1000):
+def train_logreg(X, y, learning_rate=0.01, iterations=1000, batch_size=32, init_weights=None):  
     X = np.hstack((np.ones((X.shape[0], 1)), X))  # 在特征前添加一列全为1的列  
     m, n = X.shape  
-    weights = np.zeros(n)
+    
+    # 初始化权重  
+    if init_weights is None:  
+        weights = np.random.randn(n) * 0.01  # 使用标准差为0.01的正态分布初始化权重  
+    else:  
+        weights = init_weights  
 
     for _ in range(iterations):  
-        z = np.dot(X, weights)
-        predictions = 1 / (1 + np.exp(-z))
-        gradient = np.dot(X.T, (predictions - y)) / m
+        indices = np.random.permutation(m)  # 打乱数据  
+        X_shuffled = X[indices]  
+        y_shuffled = y[indices]  
+
+        for i in range(0, m, batch_size):  # 分批处理  
+            X_batch = X_shuffled[i:i + batch_size]  
+            y_batch = y_shuffled[i:i + batch_size]  
+            
+            predictions = sigmoid(np.dot(X_batch, weights))  
+            gradient = np.dot(X_batch.T, (predictions - y_batch)) / len(y_batch)  
             weights -= learning_rate * gradient  
 
     return weights  
 
-# 对率回归预测函数  
-# 对率回归预测函数，支持多类分类
+# 逻辑回归预测函数  
 def predict_logreg(inX, weights):  
     inX = np.insert(inX, 0, 1)  # 在特征前添加1作为偏置项  
-    prob = 1 / (1 + np.exp(-np.dot(inX, weights)))
+    prob = sigmoid(np.dot(inX, weights))  
 
-    if prob < 0.5:
-        return 0
-    elif prob >= 0.5 and prob < 1.0:
-        return 1
-    else:
-        return 2  # 确保返回3个类别中的一个
+    return 1 if prob >= 0.5 else 0  # 二分类处理  
 
 # 从文件中加载数据  
 def load_data(filepath):  
@@ -59,6 +66,7 @@ def load_data(filepath):
     X = data.iloc[:, :-1].to_numpy()  # 特征  
     y = data.iloc[:, -1].to_numpy()    # 标签  
     return X, y   
+
 # 准确率计算函数  
 def calculate_accuracy(X, y, weights):  
     predictions = np.array([predict_logreg(x, weights) for x in X])  
@@ -68,16 +76,23 @@ def calculate_accuracy(X, y, weights):
 if __name__ == "__main__":  
     # 加载训练集和测试集  
     X_train, y_train = load_data("E:\\Edge站内文件下载\\病马数据集\\horse-colic.data")  
-    X_test, y_test = load_data( "E:\\Edge站内文件下载\\病马数据集\\horse-colic.test")
+    X_test, y_test = load_data("E:\\Edge站内文件下载\\病马数据集\\horse-colic.test")  
+    
+    # 数据标准化  
+    scaler = StandardScaler()  
+    X_train = scaler.fit_transform(X_train)  
+    X_test = scaler.transform(X_test)  
 
     # 训练对率回归模型  
-    weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=1000)
+    learning_rate = 0.01  
+    iterations = 1000  
+    batch_size = 32  # 随机梯度下降的小批量大小  
+    weights = train_logreg(X_train, y_train, learning_rate=learning_rate, iterations=iterations, batch_size=batch_size)  
 
     # 测试单个数据  
     y_pred = predict_logreg(X_test[11], weights)  
     print(f"预测结果: {y_pred}, 实际结果: {y_test[11]}")  
 
-
     # 计算并输出模型的准确率  
     accuracy = calculate_accuracy(X_test, y_test, weights)  
     print(f"模型的准确率: {accuracy:.4f}")