An Efficient Hybrid Algorithm for Mining Web Frequent Access Patterns

摘要

We propose an efficient hybrid algorithm WDHP in this paper for mining frequent access patterns. WDHP adopts the techniques of DHP to optimize its performance , which is using hash table to filter candidate set and trimming database. Whenever the database is trimmed to a size less than a specified threshold, the algorithm puts the database into main memory by constructing a tree, and finds frequent patterns on the tree. The experiment shows that WDHP outperform algorithm DHP and main memory based algorithm WAP in execution efficiency.