基于FCM并行算法的微博热点发现  

HOT MICROBLOGGING TOPICS DISCOVERY BASED ON PARALLEL FCM

作　　者：冯利光[1] 刘其成[1] 

机构地区：[1]烟台大学计算机与控制工程学院,山东烟台264005

出　　处：《计算机应用与软件》2015年第11期232-237,共6页Computer Applications and Software

基　　金：国家自然科学基金项目(61170224;61403329);山东省科技发展计划项目(2012GGB01017);山东省自然科学基金项目(ZR2012FL07;ZR2013FQ020)

摘　　要：微博作为发展最迅猛的信息传播平台,每天都会产生大量数据,迅速增长的数据对数据处理提出了新的要求。针对微博数据特点,对VSM模型进行改进,并在MapReduce平台上,设计微博热点话题发现并行模糊C均值(HTD-PFCM)算法。实验结果表明,基于改进VSM模型的HTD-PFCM算法具有良好的加速比,并且能够更高效地处理微博数据,发现微博热点话题。As the fastest growing information dissemination platform, microblogs produce large amounts of data every day. With the rapid growth of data, new requirement on data processing is requested. According to the characteristics of microblogging data, we improved VSM model, and designed on MapReduce platform a parallel fuzzy c-means algorithm for hot microblogging topics discovery （HTD-PFCM）. Experimental result showed that the HTD-PFCM algorithm based on the improved VSM model has a good speedup, and can handle the microblogging data more efficiently to find the hot topics of microblogs.

关 键 词：微博热点发现 并行FCM聚类算法 改进的VSM模型MapReduce框架 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]