From 153d68f1cee816926170773a8e7c6eb0d42fbfc1 Mon Sep 17 00:00:00 2001
From: 77 <3298722099@qq.com>
Date: Fri, 27 Dec 2024 14:24:45 +0800
Subject: [PATCH] =?UTF-8?q?=E6=9B=B4=E6=96=B0=20function.py?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

---
 function.py | 85 +++++++++++++++++------------------------------------
 1 file changed, 27 insertions(+), 58 deletions(-)

diff --git a/function.py b/function.py
index 95bc88e..54a7095 100644
--- a/function.py
+++ b/function.py
@@ -3,63 +3,31 @@ import numpy as np
 id = "202210423201"
 name = "陈瑞琪祥"
 
-def train_svm(X, y, learning_rate=0.01, iterations=1000):
-    """
-    支持向量机训练函数。
-    
-    Args:
-    - X (numpy array): 特征数据。
-    - y (numpy array): 标签数据。
-    - learning_rate (float): 学习率。
-    - iterations (int): 迭代次数。
-    
-    Returns:
-    - numpy array: 训练得到的权重。
-    """
-    # 初始化权重
-    weights = np.zeros(X.shape[1] + 1)  # 加1是为了偏置项
-    # 简单的梯度下降优化
-    for _ in range(iterations):
-        for i in range(len(X)):
-            if y[i] * (np.dot(X[i], weights[:-1]) + weights[-1]) < 1:
-                weights[:-1] += learning_rate * (y[i] * X[i])  # 更新权重
-                weights[-1] += learning_rate * y[i]  # 更新偏置项
-            else:
-                weights[:-1] -= learning_rate * (y[i] * X[i])  # 更新权重
-                weights[-1] -= learning_rate * y[i]  # 更新偏置项
-    
-    return weights
-
-def predict_svm(inX, weights):
-    """
-    支持向量机预测函数。
-    
-    Args:
-    - inX (numpy array): 输入待分类的样本。
-    - weights (numpy array): 训练得到的权重。
-    
-    Returns:
-    - int: 预测的分类标签。
-    """
-    # 预测标签
-    prediction = np.sign(np.dot(inX, weights[:-1]) + weights[-1])  # 加上偏置项
-    return 1 if prediction > 0 else -1
-
 def load_data(filepath):
-    """
-    从文件中加载数据。
-    
-    Args:
-    - filepath (str): 数据文件路径。
-    
-    Returns:
-    - np.array: 特征数据。
-    - np.array: 标签数据。
-    """
-    data = np.loadtxt(filepath, dtype=np.float64)
-    X = data[:, :-1]
-    y = data[:, -1]
-    return X, y
+    data = np.loadtxt(filepath, delimiter=',')
+    return data[:, :-1], data[:, -1].astype(int)
+
+def train_svm(X, y, learning_rate=0.01, iterations=1000, regularization=1.0):
+    n_samples, n_features = X.shape
+    weights = np.zeros(n_features)
+    bias = 0
+
+    for _ in range(iterations):
+        for idx, x_i in enumerate(X):
+            condition = y[idx] * (np.dot(x_i, weights) - bias) >= 1
+            if condition:
+                weights -= learning_rate * (2 * regularization * weights)
+            else:
+                weights -= learning_rate * (2 * regularization * weights - np.dot(x_i, y[idx]))
+                bias -= learning_rate * y[idx]
+
+        # Update bias
+        bias -= learning_rate * np.sum(y * (1 - y * (np.dot(X, weights) - bias)))
+
+    return weights, bias
+
+def predict_svm(inX, weights, bias):
+    return np.sign(np.dot(inX, weights) - bias)
 
 if __name__ == "__main__":
     # 加载训练集
@@ -69,8 +37,9 @@ if __name__ == "__main__":
     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=0.01, iterations=1000)
+    weights, bias = train_svm(X_train, y_train, learning_rate=0.01, iterations=1000)
 
     # 测试单个数据
-    y_pred = predict_svm(X_test[11], weights)
+    y_pred = predict_svm(X_test[11], weights, bias)
     print(f"预测结果: {y_pred}, 实际结果: {y_test[11]}")
+