更新 function.py

function.py

@@ -1,6 +1,7 @@
 import numpy as np
 from sklearn.ensemble import RandomForestClassifier
 from sklearn.preprocessing import PolynomialFeatures, StandardScaler
+from sklearn.model_selection import GridSearchCV
 
 id = "202210423104"
 name = "张钦诚"
@@ -25,39 +26,31 @@ def load_data(filepath):
 def preprocess_data(X):
     scaler = StandardScaler()
     X_standardized = scaler.fit_transform(X)
-    poly = PolynomialFeatures(degree=2, include_bias=False)
+    poly = PolynomialFeatures(degree=3, include_bias=False)  # 尝试更高的次数
     X_poly = poly.fit_transform(X_standardized)
     return X_poly
 
 def train_logreg(X, y):
-    model = RandomForestClassifier(n_estimators=113, max_depth=10, random_state=42)
-    model.fit(X, y)
-    return model
+    X_processed = preprocess_data(X)  # 预处理训练数据
+
+    # 使用网格搜索调参
+    param_grid = {
+        'n_estimators': [100, 200, 300],
+        'max_depth': [5, 10, 15],
+        'class_weight': ['balanced', None]  # 处理不平衡数据
+    }
+
+    model = RandomForestClassifier(random_state=42)
+    grid_search = GridSearchCV(model, param_grid, cv=5)
+    grid_search.fit(X_processed, y)
+    return grid_search.best_estimator_  # 返回最佳模型
 
 def predict_logreg(model, X):
-    # 确保传入的是训练好的模型
-    X_processed = preprocess_data(X.reshape(1, -1))  # 将单个样本重塑并预处理
-    return model.predict(X_processed)[0]  # 预测并返回结果
+    # 确保输入数据是一个 numpy 数组
+    if len(X.shape) == 1:
+        X = X.reshape(1, -1)  # 保持输入数据的形状为 (1, n)，符合模型的输入要求
+    X_processed = preprocess_data(X)  # 预处理测试数据
+    return model.predict(X_processed)[0]  # 确保 model 是训练好的模型
 
 if __name__ == "__main__":
-    # 加载数据
     X, y = load_data("./data/horse-colic.data")
-
-    # 划分数据集
-    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
-
-    # 数据预处理
-    X_train_processed = preprocess_data(X_train)
-    X_test_processed = preprocess_data(X_test)
-
-    # 训练模型
-    model = train_logreg(X_train_processed, y_train)
-
-    # 预测
-    y_pred = []
-    for sample in X_test:
-        y_pred.append(predict_logreg(model, sample))
-
-    # 计算正确率
-    accuracy = np.mean(y_pred == y_test)
-    print(f"测试出的正确率: {accuracy * 100:.2f}%")