关键词: 有效访问序列, 重叠, 筛选, 合并, 最大频繁项集

Abstract: The characteristics of effective access sequence in the actual application are analyzed and an efficient algorithm OUS based bottom-up strategy is proposed for mining maximal frequent itemsets.The algorithm first takes count of the browse number of each access sequence by overlapping operation，then unites and deletes the unfrequent page items according to minimum support degree given by users，afterwards sifts getting the intersections of each two user access pattern and gives birth to candidate grequent access patterns，at last，adds up the number of each candidate frequent access pattern by scanning the original database.Experimental results show that the OUS algorithm can discover user maximal frequent access patterns effectively.

Key words: effective access sequence, overlap, sift, unite, maximal frequent itemset