基于Web日志挖掘用户的浏览兴趣路径

提出一种基于Web日志发现用户浏览偏爱路径算法。引入了支持-兴趣度概念。建立用户访问矩阵以表示用户浏览网页行为,并获得访问矩阵的海明距离矩阵。通过对距离矩阵与相似度阈值的计算获得偏爱路径的候选2项子路径,再利用支持-兴趣度对候选集做进一步的过滤,最后进行合并并生产浏览偏爱路径。实验表明该算法可以有效地反映用户的浏览兴趣。

MINING USER PREFERRED BROWSING PATHS BASED ON WEB LOGS

This paper puts forward an algorithm of discovering user preferred browsing paths from Web logs.The algorithm imports support-preference concept.By creating user access matrix,user’s webpage browsing behaviour is represented,and the Hamming distance matrix of access matrix is obtained.The distance matrix and similarity threshold are computed,so the candidate preference sub-path 2-items set is gotten.According to support-preference threshold,the candidate sub-path set is further filtered.At last the preferred browsing paths are generated by merging sub-paths.Applying the algorithm in experiment,its effectiveness in reflecting users browsing interests is proved.