基于命名实体的网页推荐算法研究  

Research on Recommendation Algorithm Named Entity Based on Webpage

作　　者：李娟[1] 

机构地区：[1]宿迁市广播电视大学电子信息工程系,江苏宿迁223800

出　　处：《咸阳师范学院学报》2013年第6期40-44,共5页Journal of Xianyang Normal University

摘　　要：基于命名实体的网页推荐算法,从查询日志入手,分析用户的查询行为,给用户提供智能推荐,从而给出较好的推荐网页。提出了基于混合马尔科夫模型用于目录网页的导航链接和基于LDA特征选择的网页推荐算法用于主题网页推荐,实验结果显示,基于混合马尔科夫模型的推荐算法,达到了比较满意的效果;基于LDA特征提取的网页推荐算法优于传统的推荐算法,很好地满足了用户的需求。Webpage recommendation algorithm based on named entities, starting from the query log, query behavior analysis of users, to provide users with intelligent recommendation, so as to provide better recommendation webpage. The hybrid Markov model for directory webpage navigation links and LDA based feature selection webpage recommendation algorithm based on webpage recommendation for theme. Experimental results show that, hybrid recommendation algorithm based on Markov model, the result is satisfactory; recommendation algorithm of LDA feature extraction based on webpage recom- mendation algorithm is superior to the traditional, and is very good to meet the needs of users.

关 键 词：LDA 混合马尔科夫模型 命名实体 网页推荐 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]