基于改进的k-means算法的新闻聚类的研究  

Research on News Clustering Based on Improved K-means Algorithm

作　　者：唐啸虎 刘志锋[1,2,3] TANG Xiao-hu;LIU Zhi-feng(East China University of Technology Jiangxi Nuclear Geoscience Data Science and System Engineering Technology Research Center,Nanchang 330013,China;Engineering Research Center of Nuclear Technology Application(East China Institute of Technology),Ministry of Education,Nanchang 330013,China;East China University of Technology Information Engineering College,Nanchang 330013,China)

机构地区：[1]东华理工大学江西省核地学数据科学与系统工程技术研究中心 [2]核技术应用教育部工程研究中心 [3]东华理工大学信息工程学院,江西南昌330013

出　　处：《电脑知识与技术》2020年第10期201-203,217,共4页Computer Knowledge and Technology

基　　金：江西省核地学数据科学与系统工程技术研究中心开放基金(JETRCNGDSS201804);核技术应用教育部工程研究中心开放基金(HJSJYB2014-5)。

摘　　要：针对k-means聚类算法对初始聚类中心敏感,容易陷入局部最优的情况,提出一种改进的基于k-means算法的新闻聚类方法.在用传统k-means算法对新闻数据集进行多次聚类的基础上,使用证据累积算法对k-means算法的聚类结果进行融合,以平滑k-means算法的结果,减少波动.实验结果表明提出的方法使聚类结果的准确率从53.33%提升至77.78%.Aiming at the situation that k-means clustering algorithm is sensitive to the initial clustering center and easy to fall into local optimization,an improved news clustering method Based on k-means algorithm is proposed.Based on the traditional k-means algo rithm,the clustering results of the k-means algorithm are fused with the evidence accumulation algorithm to smooth the results of the kmeans algorithm and reduce the fluctuation.The experimental results show that the accuracy of the proposed method increases from 53.33%to 77.78%.

关 键 词：K-MEANS 新闻 聚类分析 融合 分级聚类 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]