一种从Web日志中挖掘访问模式的新算法

Algorithm of mining access patterns from Web logs

提出了一种用于从Web日志中高效挖掘Web访问模式的新算法 .设计了一个新的数据结构WAS树(WebAccessSequence树 )来记录Web访问序列和它的计数 ,避免冗长的支持度计算 ;开发出一个高效的递归算法 ,从WAS树中枚举出所有的访问模式 ,最后对算法的性能进行了分析并给出试验数据加以佐证 .这个算法的特点是不需要产生候选模式而直接挖掘出所有的访问模式 .

A data structure WAS-tree was devised to register access sequence and corresponding counts compactly and a recursive algorithm was proposed to enumerate access patterns from WAP-tree. Its advantage was that it mined all the web access patterns without generating candidate patterns. The experimental result and performance showed that the method was faster than conventional ones.