摘要
目的设计实现基于偏爱路径的个性化推荐系统原型。方法通过建立Web站点访问的一种矩阵表示模型,并据此挖掘用户浏览偏爱路径。结果分析了偏爱度与置信度的区别,提出了页面平均兴趣度的概念,改进了用户浏览偏爱路径算法。引入页面平均兴趣度的概念,给出了Web站点访问的一种矩阵表示模型,在此基础上挖掘用户浏览偏爱路径。结论该方法能准确地反映用户浏览兴趣,证明该系统具有较高的准确性。
Aim To develop a personalized recommendation system based on preferred browsing paths. Methods Preferred browsing paths could be discovered according to the computation of a URL-URL matrix matrix. Results Confidence and preference are discussed. Page average interest is presented and browsing paths mining algorithm is ameliorated. Based on these, recommendation algorithm is given and a personalized recommendation system based on preferred browsing paths is implemented. Page average is expressed by a triple group with accessing time, size of page and time, it reflects user＇s accessing way exactly. According to what we have researched, a URL-URL matrix where URL is used as rows, navigating URL as columns and page average interest as matrix elements is built. Preferred browsing paths could be discovered from the computation of this matrix. Conclusion Experiments showed that the system is effective.
出处
《西北大学学报（自然科学版）》
CAS
CSCD
北大核心
2007年第2期213-216,共4页
Journal of Northwest University（Natural Science Edition）
基金
陕西省自然科学基础研究计划基金资助项目(2006F50)
航空科学基金资助项目(2006ZC31001)
关键词
偏爱路径
个性化推荐系统
数据挖掘
页面兴趣度
preferred browsing paths
recommendation system
data minding
page average interest
