更新 function.py

function.py

@@ -28,7 +28,7 @@ name = "邓龙树"
 
 
 # 这里写自己的代码实现
-def train_svm(X, y, learning_rate=0.001, iterations=1000):
+def train_svm(X, y, learning_rate=0.01, iterations=1000 ):
     """
     支持向量机训练函数。
 
@@ -40,19 +40,16 @@ def train_svm(X, y, learning_rate=0.001, iterations=1000):
     Returns:
     - numpy array: 训练得到的权重。
     """
-    n_samples, n_features = X.shape
-    # 初始化权重向量
-    weights = np.zeros(n_features)
-    eta = 0.001  # 学习率
-    epochs = 1000  # 迭代次数
-
-    for _ in range(epochs):
-        for idx, x_i in enumerate(X):
-            condition = y[idx] * (np.dot(x_i, weights))
-            if condition < 1:
-                weights = weights + eta * ( (y[idx] * x_i) - (2  / (n_samples * epochs)) * weights)
-            else:
-                weights = weights + eta * (- (2  / (n_samples * epochs)) * weights)
+    # 这里采用简单的梯度下降法示例来训练线性SVM（实际应用中可使用更高效的SMO等算法优化）
+    num_samples, num_features = X.shape
+    learning_rate = 0.01  # 学习率，可作为参数调整
+    iterations = 1000  # 迭代次数，可作为参数调整
+    weights = np.zeros(num_features)
+    # 迭代更新权重
+    for _ in range(iterations):
+        for i in range(num_samples):
+            if y[i] * (np.dot(X[i], weights)) < 1:
+                weights += learning_rate * (y[i] * X[i])
     return weights
 
 
@@ -67,8 +64,14 @@ def predict_svm(inX, weights):
     Returns:
     - int: 预测的分类标签。
     """
-    prediction = np.sign(np.dot(inX, weights))
-    return prediction if prediction!= 0 else -1
+    # 计算决策函数值（这里是线性SVM的简单情况）
+    result = np.dot(inX, weights)
+    # 根据结果判断类别（简单的多分类策略示例，可优化）
+    if result >= 0.5:
+        return 1
+    elif result < -0.5:
+        return 3
+    return 2
 
 
 def load_data(filepath):
@@ -81,38 +84,46 @@ def load_data(filepath):
     Returns:
     - np.array: 特征数据。
     - np.array: 标签数据。
+
     """
     data = []
     labels = []
-    with open(filepath, 'r') as file:
-         lines = file.readlines()
-         for line in lines:
-            try:
-                line = line.strip().split()
-                if '?' not in line:  # 只处理不包含缺失值的行
-                    features = [float(x) for x in line[:-1]]
-                    label = float(line[-1])
-                    if label == 0:
-                        label = -1  # 将类别0转换为 -1 ，符合SVM二分类的约定
-                    data.append(features)
-                    labels.append(label)
-            except ValueError:
-                print(f"数据解析出现问题，跳过该行数据: {line}")
-                continue
-    return np.array(data), np.array(labels)
+    with open(filepath, 'r') as f:
+        for line in f.readlines():
+            parts = line.strip().split()
+            features = []
+            for p in parts[:-1]:
+                try:
+                    features.append(float(p))
+                except ValueError:
+                    features.append(0)  # 将无法转换的字符替换为0，也可以用其他合适值
+            label = int(parts[-1])
+            data.append(features)
+            labels.append(label)
+    data = np.array(data)
+    labels = np.array(labels)
+    # 选择特征23（从1开始由左向右数）作为目标变量
+    data = data[:, 22].reshape(-1, 1)  # 修正这里的语法错误，将--1改为-1
+    return data, labels
+
 
 if __name__ == "__main__":
     # 加载训练集
     X_train, y_train = load_data("./data/horse-colic.data")
-
     # 加载测试集
     X_test, y_test = load_data("./data/horse-colic.test")
 
     # 训练支持向量机模型
-    weights = train_svm(X_train, y_train, learning_rate=0.001, iterations=1000)
-
-    # 测试单个数据
+    weights = train_svm(X_train, y_train, learning_rate=0.01, iterations=1000)
 
+    # 统计正确分类的样本个数
+    correct_count = 0
     for i in range(len(X_test)):
         y_pred = predict_svm(X_test[i], weights)
-        print(f"预测结果: {y_pred}, 实际结果: {y_test[i]}")
+        if y_pred == y_test[i]:
+            correct_count += 1
+
+    # 计算准确率（得分）
+    score = correct_count / len(X_test) * 100
+
+    print(f"预测准确率: {score}%")