关键词: Web挖掘, 推荐系统, 关联规则, 序列模式, 聚类算法, AIR算法

Abstract: Internet users often spend much time finding useful pages.Recommendation system does such a job that it can help user locate information and increase the users’ loyalty.In respect that web mining is good at dealing with massive data and sparse data，web mining is widely used in recommendation system.Most important technologies in web mining include clustering，association rules，sequence pattern.Howerver，these technologies have a limitation that it is difficult to strike the balance of precision and coverage.In this paper，we integrate these technologies and raise a new algorithm（AIR）.Our experiment data，performed on real usage data，indicate that AIR can achieve dramatic improvements in terms of recommendation effectiveness.

Key words: Web mining, recommendation system, association rules, sequence pattern, clustering, AIR