更新 function.py

function.py

@@ -7,7 +7,7 @@ from sklearn.preprocessing import StandardScaler
 id = "202210423201"
 name = "陈瑞琪祥"
 
-
+import numpy as np
 
 # 数据加载函数
 def load_data(filepath):
@@ -31,20 +31,26 @@ def load_data(filepath):
     
     return np.array(data), np.array(labels)
 
+# 线性SVM训练函数
 def train_svm(X, y):
-    """
+    n_samples, n_features = X.shape
-    支持向量机训练函数。
+    weights = np.zeros(n_features)
-    """
+    bias = 0
-    scaler = StandardScaler()
+    learning_rate = 0.01
-    X_scaled = scaler.fit_transform(X)
+    iterations = 1000
     
-    model = SVC(kernel='rbf', C=1.0, gamma='scale', random_state=42)
+    for _ in range(iterations):
-    model.fit(X_scaled, y)
+        for idx, x_i in enumerate(X):
-    return model, scaler
+            condition = y[idx] * (np.dot(x_i, weights) - bias) >= 1
+            if not condition:
+                weights -= learning_rate * (2 * weights - np.dot(x_i, y[idx]))
+                bias -= learning_rate * y[idx]
     
-def predict_svm(model, scaler, X):
+    return weights, bias
-    X_scaled = scaler.transform(X)
+
-    return model.predict(X_scaled)
+# 预测函数
+def predict_svm(X, weights, bias):
+    return np.sign(np.dot(X, weights) - bias)
 
 if __name__ == "__main__":
     # 加载数据
@@ -54,10 +60,10 @@ if __name__ == "__main__":
     X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
     
     # 训练SVM模型
-    svm_model, scaler = train_svm(X_train, y_train)
+    weights, bias = train_svm(X_train, y_train)
     
     # 预测测试集
-    y_pred = predict_svm(svm_model, scaler, X_test)
+    y_pred = predict_svm(X_test, weights, bias)
     
     # 评估模型
     accuracy = accuracy_score(y_test, y_pred)
@@ -65,5 +71,5 @@ if __name__ == "__main__":
     
     # 测试单个数据
     test_sample = X_test[11]  # 选择第12个样本（从0开始计数）
-    y_pred_single = predict_svm(svm_model, scaler, test_sample.reshape(1, -1))
+    y_pred_single = predict_svm(test_sample.reshape(1, -1), weights, bias)
     print(f"\n单个样本预测结果: {y_pred_single[0]}, 实际结果: {y_test[11]}")