K-means聚类算法的研究和应用  

Research and Application of K-means Clustering Algorithm

作　　者：熊志斌[1] 朱剑锋[1] 王冬[1] 

机构地区：[1]琼州学院电子信息工程学院,海南三亚572022

出　　处：《电脑编程技巧与维护》2014年第8期10-12,共3页Computer Programming Skills & Maintenance

摘　　要：介绍了K-means算法的思想,分析了在文档聚类中运用K-means算法的步骤。以开源的机器学习软件Weka为平台,详细论述在Weka上进行文档聚类的前端处理过程,利用搜狗语料库中的文档在Weka上进行了Kmeans算法的聚类测试。实验结果表明,K-means算法在Web文档聚类中表现出较好的效果。根据实验结果,分析了K-means算法存在的不足和聚类分析中特征选择的重要性。This paper introduced the principal of the K-means algorithm, and presented the method by which K-means algorithm is applied to document clustering. The front-end process of documents clustering was discussed in detail, based on Weka that is an open source machine learning software. The K-means clustering algorithm was tested on Weka with documents corpus from Sogou Corporation, and results from experiments show that the K-means algorithm is effective in Web document clustering. According to experimental data, the defect of K-means algorithm was discussed, and importance of feature selection was discussed in unsupervised clustering.

关 键 词：K-MEANS算法 聚类分析 WEB文档 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]