更新 function.py

function.py

@@ -27,115 +27,52 @@ name = "邓龙树"
 # user_code.py
 
 import numpy as np
+from sklearn.preprocessing import MinMaxScaler
 
-from sklearn.preprocessing import StandardScaler
-
-def sigmoid(z):
-    """计算 sigmoid 函数"""
-    return 1 / (1 + np.exp(-z))
-
-def train_logreg(X, y, learning_rate=0.01, iterations=1000, l2_penalty=0.01):
-    """对率回归训练函数，包括 L2 正则化。"""
-    weights = np.zeros(X.shape[1])
-    
+def train_logreg(X, y, learning_rate=0.01, iterations=1000):
+    num_samples, num_features = X.shape
+    weights = np.zeros(num_features)
     for _ in range(iterations):
         z = np.dot(X, weights)
-        predictions = sigmoid(z)
-        gradient = np.dot(X.T, (predictions - y)) / y.size + l2_penalty * weights  # 添加 L2 正则化项
+        h = 1 / (1 + np.exp(-z))
+        gradient = np.dot(X.T, (h - y)) / num_samples
         weights -= learning_rate * gradient
-    
     return weights
 
-def predict_logreg(X, weights):
-    """使用训练好的权重预测标签"""
-    z = np.dot(X, weights)
-    predictions = sigmoid(z)
-    return np.where(predictions >= 0.5, 1, 0)
+def predict_logreg(inX, weights):
+    z = np.dot(inX, weights)
+    return 1 if 1 / (1 + np.exp(-z)) > 0.5 else 0
 
 def load_data(filepath):
-    """从文件中加载数据，处理不规则列数问题"""
-    with open(filepath, 'r') as file:
-        lines = file.readlines()
-    
-    max_columns = 0
-    data = []
-    
-    for line in lines:
-        parts = line.strip().split()
-        max_columns = max(max_columns, len(parts))
-        data.append(parts)
-    
-    # 填充缺失的列
-    for i in range(len(data)):
-        if len(data[i]) < max_columns:
-            data[i].extend(['0'] * (max_columns - len(data[i])))  # 使用'0'填充缺失列
-
-    # 转换成 numpy 数组，处理缺失值
-    data_array = []
-    for row in data:
-        cleaned_row = []
-        for value in row:
-            if value == '?':  # 替换无法转换的值
-                cleaned_row.append(0)  # 使用 0 填充
-            else:
-                cleaned_row.append(float(value))  # 转换为 float
-        data_array.append(cleaned_row)
-
-    data_array = np.array(data_array)
-
-    return data_array[:, :-1], data_array[:, -1]
-
-def preprocess_data(X, y):
-    """数据预处理，合并特征 23 的值，并删除特征 3 和特征 24~28，以及使用库函数进行特征缩放和归一化。"""
-    # 合并特征 23 的值和删除特征
-    X, y = preprocess_data_base(X, y)
-
-    # 选择特征 23 作为目标变量，将其取值 2（死亡）和 3（安乐死）合并为一个值 0（死亡）
-    y = np.where(y == 2, 0, y)
-    y = np.where(y == 3, 0, y)
-
-    # 使用库函数进行标准化处理
-    scaler = StandardScaler()
-    X = scaler.fit_transform(X)
-
+    data = np.genfromtxt(filepath, delimiter=',', dtype=np.float64)
+    # 处理缺失值
+    data[np.isnan(data)] = 0
+    X = data[:, :-1]
+    y = data[:, -1]
+    # 处理特征 23，合并取值 2 和 3 为 0
+    y[y == 2] = 0
+    y[y == 3] = 0
+    # 去掉特征 3、特征 24~28
+    X = np.delete(X, [2, 23], axis=1)
     return X, y
 
-def preprocess_data_base(X, y):
-    """数据预处理，删除特征 3 和特征 24~28。"""
-    # 删除特征 3 和特征 24~28（假设特征索引从 0 开始）
-    X = np.delete(X, [2] + list(range(23, 28)), axis=1)
-
-    return X, y
-
-def evaluate_model(X_test, y_test, weights):
-    """评估模型准确率"""
-    predictions = predict_logreg(X_test, weights)
-    accuracy = np.mean(predictions == y_test)
-    return accuracy
-
 if __name__ == "__main__":
-    # 加载数据
-    X, y = load_data("./data/horse-colic.data")
+    # 加载训练集
+    X_train, y_train = load_data("./data/horse-colic.data")
 
-    # 数据预处理
-    X, y = preprocess_data(X, y)
+    # 对训练集特征进行归一化
+    scaler = MinMaxScaler()
+    X_train = scaler.fit_transform(X_train)
 
-    # 尝试不同的正则化强度
-    l2_penalties = [0.01, 0.1, 1.0]
-    
-    for l2_penalty in l2_penalties:
-        # 训练模型
-        weights = train_logreg(X, y, learning_rate=0.01, l2_penalty=l2_penalty)
-
-        # 加载测试数据并进行预处理
+    # 加载测试集
     X_test, y_test = load_data("./data/horse-colic.test")
-        X_test, y_test = preprocess_data(X_test, y_test)
 
-        # 评估模型准确率
-        accuracy = evaluate_model(X_test, y_test, weights)
-        print(f"L2 正则化强度: {l2_penalty}, 准确率: {accuracy:.4f}")
+    # 对测试集特征进行归一化
+    X_test = scaler.transform(X_test)
 
-    # 打印预测结果和实际结果（以最后一个正则化强度为例）
-    predictions = predict_logreg(X_test, weights)
-    for i in range(min(10, len(y_test))):
-        print(f"预测结果: {predictions[i]}, 实际结果: {y_test[i]}")
+    # 训练对率回归模型
+    weights = train_logreg(X_train, y_train, learning_rate=0.01, iterations=1000)
+
+    # 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]}")