7777

function.py

@@ -1,6 +1,7 @@
 import numpy as np
 from sklearn.preprocessing import StandardScaler
 from sklearn.linear_model import LogisticRegression
+from sklearn.model_selection import train_test_split, GridSearchCV
 
 """
 实验说明:
@@ -25,7 +26,6 @@ id = "202210423221"
 name = "廖子豪"
 
 def load_data(filepath):
-    # 使用 numpy 的 genfromtxt 函数加载数据，自动处理缺失值
     data = np.genfromtxt(filepath, delimiter=' ', missing_values='?', filling_values=0)
     labels = data[:, -1].astype(int)
     features = data[:, :-1]
@@ -33,30 +33,28 @@ def load_data(filepath):
     # 将标签转换为二分类问题
     labels = np.where((labels == 2) | (labels == 3), 1, 0)
     
-    # 删除不相关特征（如果需要）
-    # features = np.delete(features, [2] + list(range(23, 28)), axis=1)
-    
     # 数据标准化
     scaler = StandardScaler()
     features = scaler.fit_transform(features)
     
-    return features, labels
+    return features, labels, scaler
 
 def train_logreg(X, y, random_state=42):
-    # 创建逻辑回归模型，使用liblinear求解器
+    # 使用 GridSearchCV 进行参数调优
+    param_grid = {'C': [0.01, 0.1, 1, 10, 100], 'penalty': ['l1', 'l2']}
     model = LogisticRegression(solver='liblinear', random_state=random_state)
-    model.fit(X, y)
+    grid_search = GridSearchCV(model, param_grid, cv=5)
-    return model
+    grid_search.fit(X, y)
+    return grid_search.best_estimator_
 
 def predict_logreg(inX, model):
-    # 使用模型进行预测
     return int(model.predict([inX])[0])
 
 # 示例使用
 if __name__ == "__main__":
     # 加载训练数据和测试数据
-    X_train, y_train = load_data("./data/horse-colic.data")
+    X_train, y_train, scaler = load_data("./data/horse-colic.data")
-    X_test, y_test = load_data("./data/horse-colic.test")
+    X_test, y_test, _ = load_data("./data/horse-colic.test")  # 使用相同的scaler
 
     # 训练模型
     model = train_logreg(X_train, y_train, random_state=42)