摘要
基于链接结构的PageRank排序算法,存在平均分配权值、主题漂移、忽视用户兴趣等问题。针对这些问题,结合搜索词与网页关键词以及用户搜索历史与网页关键词之间的相似度,提出基于主题相关度与用户历史的PageRank改进算法THPR。通过查询词与网页关键词的相关性分析解决网页的相关程度,通过对用户历史搜索记录的分析,在算法中增加用户兴趣度,使PR值分配更为合理。仿真结果表明,THPR算法与PageRank算法相比,查准率与用户搜索满意度有明显提升。
PageRank algorithm based on link structure has some defects such as drifting theme and ignoring interest.Aiming at this kind of problem,considering the similarity between the search term and the web keywords,and the similarity between user search history and page keywords,the improved algorithm THPR algorithm was proposed,combining with keyword correlation and user history search.The degree of the web page was solved by the correlation analysis between the query terms and the keyword of the web page.By analyzing user history search records,user interest was added in the algorithm,so that the distribution of PR values was more reasonable.The simulation results show that the precision and user search satisfaction of THPR algorithm are significantly improved compared with PageRank algorithm.
作者
林婷薇
莫路锋
薛晨杰
LIN Ting-wei;MO Lu-feng;XUE Chen-jie(College of Information Engineering,Zhejiang Agriculture and Forestry University,Hangzhou 311300,China)
出处
《计算机工程与设计》
北大核心
2019年第8期2265-2269,2277,共6页
Computer Engineering and Design
