一种预测群体用户访问行为的算法

An Algorithm of Predicting Users Access Behaviors

在电子商务发展中,商家需要理解用户访问网站的行为,为用户提供个性化服务,从而吸引用户购买商品。挖掘用户访问网站的行为是商家一个急需解决的问题,通过对Web日志进行挖掘是解决该问题的重要研究方法。提出了网页兴趣信息素的新概念,它是由页面相对浏览时间和点击率构建而成,利用兴趣信息素设计了基于蚁群算法的群体用户访问路径挖掘算法,根据挖掘结果预测用户访问行为。实验结果表明,兴趣信息素可以有效地预测用户的兴趣变化,能准确地反映用户访问模式,提高了预测群体用户访问行为的准确率。

During the developing of e-commerce, enterprises should understand the behaviors of users browsing the Website, offer personal services to those that be leaded to buy their commodities. Mining user access behaviors is one problem that is badly in need of being solved by businessmen, while Web log mining suggests an important research method. In this paper, put forward a new concept of Web page interest pheromone. Both text relative browsing time and click rate define Web page interest pheromone. Group user access path mining algorithm is based on ant colony algorithm, interest pheromone being involved in designing. And according to mining results, user access behaviors are predicted. Experimental results show interest pheromone can efficiently forecast how user interests change, accurately reflect user access mode and improve the predicting accuracy of group user access behaviors.