更新 function.py

function.py

@@ -43,79 +43,52 @@ name = "方子涵"
 # 这里写自己的代码实现
 import numpy as np
 
-"""
-请在这里填写学号和姓名，不填写即成绩无效。
-"""
-id = "202210423220"
-name = "方子涵"
-
-
 def euclidean_distance(inX, dataSet):
-    """
-    计算待分类样本与数据集中每个样本的欧式距离。
+    return np.sqrt(((inX - dataSet) ** 2).sum(axis=1))
 
-    Args:
-    inX (numpy array): 输入待分类的样本。
-    dataSet (numpy array): 训练数据集。
-
-    Returns:
-    numpy array: 每个样本与待分类样本之间的欧式距离。
-    """
-    # 计算差值
-    diff = dataSet - inX
-    # 计算平方
-    sq_diff = diff ** 2
-    # 计算平方和
-    sq_distance = sq_diff.sum(axis=1)
-    # 开方
-    distance = np.sqrt(sq_distance)
-    return distance
-
-
-def classify0(inX, dataSet, labels, k=3):
-    """
-    使用欧式距离的KNN分类算法。
-
-    Args:
-    inX (numpy array): 输入待分类的样本。
-    dataSet (numpy array): 训练数据集。
-    labels (list or numpy array): 训练集对应的标签。
-    k (int): kNN 算法参数。
-
-    Returns:
-    int: 预测的分类标签。
-    """
+def weighted_knn(inX, dataSet, labels, k=3):
     # 计算欧式距离
     distances = euclidean_distance(inX, dataSet)
     
-    # 获取距离最近的k个样本的索引
-    k_indices = np.argsort(distances)[:k]
+    # 获取距离和对应的索引
+    sorted_distances_idx = np.argsort(distances)
     
-    # 获取这些样本对应的标签
-    k_nearest_labels = labels[k_indices]
+    # 获取k个最近邻的标签和对应的距离
+    k_nearest_labels = labels[sorted_distances_idx[:k]]
+    k_nearest_distances = distances[sorted_distances_idx[:k]]
     
-    # 统计每个标签出现的次数
-    label_counts = np.bincount(k_nearest_labels)
+    # 创建权重数组，距离越近权重越大
+    weights = 1 / k_nearest_distances
     
-    # 返回出现次数最多的标签
-    return np.argmax(label_counts)
+    # 归一化权重
+    weights = weights / weights.sum()
     
+    # 计算加权投票
+    label_prob = np.dot(weights, k_nearest_labels)
     
-def load_data(file_path):
-    """
-    加载 TSV 格式数据文件。
-    """
-    data = np.loadtxt(file_path, delimiter="\t")
-    features = data[:, :-1]
-    labels = data[:, -1].astype(int)
-    return features, labels
+    # 返回概率最高的标签
+    return np.argmax(label_prob)
 
+def normalize_data(dataSet):
+    minVals = dataSet.min(0)
+    maxVals = dataSet.max(0)
+    ranges = maxVals - minVals
+    normDataSet = np.zeros_like(dataSet)
+    m = dataSet.shape[0]
+    normDataSet = dataSet - minVals
+    normDataSet = normDataSet / ranges
+    return normDataSet, ranges, minVals
+
+def classify0(inX, dataSet, labels, k=3):
+    # 特征缩放
+    norm_dataSet, _, _ = normalize_data(dataSet)
+    norm_inX = (inX - norm_dataSet.min(0)) / (norm_dataSet.max(0) - norm_dataSet.min(0))
+    
+    # 使用加权KNN进行分类
+    return weighted_knn(norm_inX, norm_dataSet, labels, k)
 
-if __name__ == "__main__":
 # 以下代码可用于测试，可自行随意修改
-
+if __name__ == "__main__":
     train_file = "datingTrainSet2.tsv"
-
     train_data, train_labels = load_data(train_file)
-
     print(classify0([65915, 3.868359, 0.128078], train_data, train_labels))