欧式距离与标准化欧式距离在k近邻算法中的比较

Comparison Between Euclidean Distance and Standardized Euclidean Distance in K-nearest Neighbor Algorithm

相似性度量是综合评定两个数据样本之间差异的指标,欧式距离是较为常用的相似性度量方法之一。本文分析了欧式距离与标准化的欧式距离在KNN算法中对数据分类的影响。仿真实验结果表明,当向量之间的各维度的尺度差别较大时,标准化的欧式距离较好地改善了分类的性能。

Similarity measurement is an index to evaluate the difference between two data samples.Euclidean distance is one of the most common similarity measurement methods.This paper analyzes the influence of Euclidean distance and standardized Euclidean distance on data classification in KNN algorithm.The simulation results show that the normalized Euclidean distance improves the classification performance as the scales of the dimensions between vectors differ greatly.