基于用户多页面浏览模式下的网络结构推荐系统的研究

摘要

针对用户普遍使用的多页面浏览器产生树型结构的浏览路径，Web日志中将会呈现非时序的日志记录。本文提出了一种新的自上而下的用户访问路径收集算法，进而得出的用户在一次会话中可能访问的复数目的页面，由此得出全局目的页面访问频度矩阵，此矩阵的数据作为实现基于网络结构的推荐系统的核心数据。%In view of browsing path for the tree structure of multi page browser which is widely used by users, the Web log will show non sequential log records. This paper presents a new top⁃down user access path collection algorithm, and then come to the complex page a user in a session may visit, resulting in a global page access frequency matrix. This matrix data could be used as core data based on the recommendation system from the network structure.