AlgorithminWebTopicDetection.PDFVIP

International Conference on Advanced Information and Communication Technology for Education (ICAICTE 2014) A Comparative Study between Single-Pass Algorithm and K-means Algorithm in Web Topic Detection Xi Ting 1,a , Li Jufang 2,b 1National University of Defense Technology，Changsha，Hunan，China 2National University of Defense Technology， Changsha ， Hunan， China axtbeihang@, b lijufang@, *Corresponding author :Xi Ting Abstract. As with the extensive application of the Internet, the explosive growth of information and unprecedented enthusiasm of users, the monitoring and management of Web content is becoming more and more imminent. Although traditional Single-Pass algorithm and K-means algorithm each has shortcomings, they are widely used in clustering analysis because of their relatively simple prin- ciples and fast computing speed. This paper firstly describes the overall flow of the entire topic of detection, then we make a comparison between Single-Pass algorithm and K-means algorithm. In order to verify the comparison, finally, an experiment is designed. The result shows that Single-Pass algorithm is better than K-means algorithm in Web topic detection. Keywords: Text data, Clustering algorithm, Topic detection. Single-Pass算法和K-means算法在Web话题分析中的对比性研究 1, a 2,b,* 习婷 , 李菊芳 1 国防科学技术大学信息系统与管理学院，长沙，湖南，中国 2 国防科学技术大学信息系统与管理学院，长沙，湖南，中国 axtbeihang@, b lijufang@, * 习婷 中文摘要. 互联网的广泛应用、信息的爆 1. 引言 炸式增长以及网民参与热情的空前高涨， 使得对Web 内容的监控和管理迫在眉睫。 Internet 已经成为继报纸、电视、广播 传统的Single-Pass 算法和K-means 算法虽