一种有效的时序传递模式挖掘算法

Most existing web association mining algorithms aim to mine frequent access patterns ignoring an important relationship, i.e. indirect association, among the accessed webs. The indirect association can discover an interesting web pattern, where some items are infrequent in the whole dataset but may be frequent via a bridging mediator item in a particular time period. More difficult, such an indirect pattern will vary with time. In this study, a novel approach, namely TIFP-Mine, to mining temporal indirect association rules from temporal web data is proposed. The proposed algorithm can capture all temporal transition patterns and relevant extended patterns. Experimental results confirm that TIFP-mine algorithm is efficient and effective.