A Method of Predicting Users' Behaviors Based on Inter-transaction Association Rules
Yanyu Zhang, Yonggong Ren · 2009
Association rules is one of web data mining methods, taking advantage of the knowledge acquired through the web log and finding the user's navigational behavior. Recently, nearly all researches of association rules are based on intra-transaction. They all focus on the relationship among web pages. But user is the center of all the internet services, and they should be given more considered. In this paper, a new method of predicting users` behaviors based on inter-transaction association rules is proposed. Through the improved Mafia algorithm, the maximum frequent itemsets with CUI can be found. We generate the inter-transaction association rules, discovering the relationship among users, predicting next pages the user visited. Experimental results prove that this method provides more accurate prediction results than former researches, and users will get more of the content they want.