用户浏览偏爱模式挖掘算法的研究

Research on the Algorithm of Mining Preferred Navigation Patterns

针对当前的挖掘算法只是简单地把频繁访问路径作为用户浏览的兴趣路径的问题，充分地考虑了用户在页面上的浏览时间和在路径选择上表现出来的浏览偏爱，提出了基于远程代理数据收集的浏览偏爱模式挖掘算法．该算法先利用客户端的远程代理收集用户浏览信息，然后划分成用户事务，最后利用一个递归过程找出用户浏览偏爱模式．实验证明：该算法比当前的频繁访问路径算法在用户浏览兴趣度量上更准确．

The traditional algorithms simply regard frequent access paths as interesting navigation paths. It's not accurate. Considering user navigation interest by page navigating time and path choosing patterns, the concept of navigation preference and the algorithm for mining preferred navigation patterns are proposed. User navigation information is first collected from the remote agent and then is divided into user session. Then, the preferred navigation paths are mined by a recursion procedure. Experiments show our algorithm is more accurate than the other traditional algorithms of frequent access paths in measuring user navigation interests.