chore: only use numpy

function.py

@@ -52,31 +52,23 @@ def classify0(inX, dataSet, labels):
     Returns:
     - int: 预测的分类标签。
     """
-    # 获取数据集的大小
-    dataSetSize = dataSet.shape[0]
-
     # 计算数据集的协方差矩阵的逆
     VI = np.linalg.inv(np.cov(dataSet.T))
 
-    # 计算 Mahalanobis 距离
+    # 计算 inX 与数据集的差值
-    distances = np.zeros(dataSetSize)
+    diff = dataSet - inX
-    for i in range(dataSetSize):
+
-        distances[i] = mahalanobis(inX, dataSet[i], VI)
+    # 使用矩阵运算计算马氏距离
+    distances = np.sqrt(np.sum(diff @ VI * diff, axis=1))
 
     # 对距离排序，获取前 k 个最近的样本索引
-    sortedDistIndices = distances.argsort()
+    sortedDistIndices = np.argsort(distances)
 
     # 统计前 k 个样本的标签
-    classCount = {}
+    classCount = np.bincount(labels[sortedDistIndices[:8]])
-    for i in range(9):
-        voteLabel = labels[sortedDistIndices[i]]
-        classCount[voteLabel] = classCount.get(voteLabel, 0) + 1
 
     # 返回出现次数最多的标签
-    sortedClassCount = sorted(
+    return np.argmax(classCount)
-        classCount.items(), key=operator.itemgetter(1), reverse=True
-    )
-    return sortedClassCount[0][0]
 
 
 def load_data(file_path):

requirements.txt

@@ -1,3 +1,2 @@
 numpy==2.1.1
 requests==2.32.3
-scipy