一种基于最大频繁项目集的挖掘事务间关联规则方法

Method of Mining Inter-transaction Association Rules Based on Maximum Frequent Itemsets

Web事务间关联规则挖掘是通过发现网页之间的关联关系来预测用户的兴趣。提出一种新的事务间关联规则挖掘方法,通过对MAFIA算法改进,得到最大频繁项目集的同时得到对应的共有用户集,通过对事务内到事务间最大频繁项目集的转换,分析不同用户之间的关系,分析用户对网站上不同网页的访问数据,直接发现不同用户之间的关联关系来预测用户的兴趣。该方法经试验证明能够更加全面的预测用户感兴趣的网页,更好地为用户提供个性化服务。

Mining Web inter-transaction association rules is to predict the users＇ interesting by finding the relationship among Web pages. We proposed a new method of mining inter-transaction association rules, that improve the Mafia for getting the maximum frequent itemsets and the relative CUI（Common User Intersection）, transform the maximum fre- quent itemsets from intra-transaction into inter-transaction, analysed the relationship among different users and analysis the visiting information of Web pages. The association rules among different users instead of among different Web pages can he found. The experiment proved that this method provides more Web pages which could content users＇ interesing and serve the users more personalized.