摘要
用户兴趣的度量和用户兴趣的修正是个性化服务研究的重要内容。本文以用户最小浏览行为组合为基础,通过引入页面浏览率,改进页面驻留时间的计算方法,建立以页面浏览率、驻留时间和浏览速度为变量的兴趣度估计函数,提高用户兴趣度估计的准确性。同时,本文还将兴趣度导入到向量空间模型,采用二层树状结构表示用户兴趣,并提出用户兴趣定期修正方法,以缓解用户兴趣实时修正带来系统性能的下降。
The calculation of user's interest degree and amendment of user's interest are the essential content of personalized services research. Based on a set of user browsing behaviors, it introduced the concept of browsing rate, improve measurement method of stay time on webpage and established a calculation function of interest degree with variables browsing rate, stay time and browsing speed of webpage. At the same time, interest degree was imported into Vector Space Model and user's interest was pre- sented by two- layer tree structure on the paper. Besides, fixed period araendment method of user's interest was given on the pa- per at last which could not only modify user interest in time, but also ease the problem of system's performance drop caused by user interest' real - time updating.
出处
《现代情报》
CSSCI
2014年第1期46-48,55,共4页
Journal of Modern Information
基金
江苏省高校哲学社会科学研究重大项目与重点项目"SSME视角下江苏省创新型电子服务模式与对策研究"(项目编号:2012ZDIXM022)
关键词
用户浏览行为
兴趣度计算
用户兴趣修正
向量空间模型
user's browsing behaviors
interest degree calculation
user's interest amendment
vector space model