From 16f4b687160d3535efb4a4cc56c2318f1e5ab57e Mon Sep 17 00:00:00 2001
From: Dengddd <3149942468@qq.com>
Date: Tue, 22 Oct 2024 20:01:32 +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 | 24 +++++++++++++++++++-----
 1 file changed, 19 insertions(+), 5 deletions(-)

diff --git a/function.py b/function.py
index 07bed3a..4cb1f3f 100644
--- a/function.py
+++ b/function.py
@@ -45,18 +45,32 @@ import numpy as np
 import os
 from collections import Counter
 
-def classify(input_vector, dataSet, labels, k=3):  # 设置默认值 k=1
-    """使用 KNN 算法对输入样本进行分类。"""
-    # 计算输入向量与训练集之间的欧氏距离
-    distances = np.linalg.norm(dataSet - input_vector, axis=1)
+def classify_mahalanobis(input_vector, dataSet, labels, k=1):
+    """使用 KNN 算法并基于马氏距离对输入样本进行分类。"""
+    # 计算数据集的协方差矩阵及其逆矩阵
+    cov_matrix = np.cov(dataSet, rowvar=False)
+    
+    # 对协方差矩阵添加正则化
+    cov_matrix += np.eye(cov_matrix.shape[0]) * 1e-5
+    
+    # 计算协方差矩阵的逆矩阵
+    inv_cov_matrix = np.linalg.inv(cov_matrix)
+    
+    # 计算马氏距离
+    diffMat = dataSet - input_vector
+    distances = np.sqrt(np.sum(np.dot(diffMat, inv_cov_matrix) * diffMat, axis=1))
+
     # 根据距离排序并获取索引
     sorted_indices = distances.argsort()
+    
     # 选择距离最近的k个点
     k_nearest_labels = labels[sorted_indices[:k]]
+    
     # 统计k个点中各类别的数量
     count = Counter(k_nearest_labels)
+    
     # 返回数量最多的类别标签
-    return count.most_common(1)[0][0]
+return count.most_common(1)[0][0]
 
 def load_dataset(directory):
     """加载数据集中的图像，并将其转换为特征向量和标签。"""