您的位置:首页  > 论文页面

基于欧氏距离的K-means算法优化

发表时间:2019-12-31  浏览量:423  下载量:39
全部作者: 李轮,宋文广,沈翀,张伟委,邓健
作者单位: 长江大学计算机科学学院
摘 要: 传统K-means聚类算法在使用上有太多的局限性。针对K-means算法,在基于欧氏距离相似度计算的基础上,利用现有的一些算法,从聚类值k大小的确定和初始聚类中心的选取这两方面进行相应的优化。通过Matlab工具进行数据测试实验,采用轮廓系数衡量算法结果,将优化前及优化后算法从不同方面作比较并分析结果。实验证明该算法优化后的表现更佳。
关 键 词: 计算机软件;聚类分析;K-means;相似度计算;K-means++;轮廓系数
Title: Optimization of K-means algorithm based on Euclidean distance
Author: LI Lun, SONG Wenguang, SHEN Chong, ZHANG Weiwei, DENG Jian
Organization: School of Computer Science, Yangtze University
Abstract: For the traditional K-means clustering algorithm, there are too many limitations in its use. For the K-means algorithm, based on Euclidean distance similarity calculation, some existing algorithms are used to optimize the judgment of clustering value k and the selection of individual initial clustering centers. Data test experiments are carried out with Matlab tools, and the results of the algorithm are measured with silhouette coefficient. The results are compared from different aspects before and after optimization. It is proved that the optimized algorithm performs better.
Key words: computer software; cluster analysis; K-means; similarity calculation; K-means++; silhouette coefficient
发表期数: 2019年12月第6期
引用格式: 李轮,宋文广,沈翀,等. 基于欧氏距离的K-means算法优化[J]. 中国科技论文在线精品论文,2019,12(6):889-895.
 
0 评论数 0
暂无评论
友情链接