Abstract: To find more information about user interest on hierarchical structure website，this paper proposes a novel efficient method，which is used for providing dynamic multi-level user interests.The method using the hierarchy characteristics of website and user interest，based on ant colony algorithm，considering a user’s website browsing process as an ant foraging behavior，counts degrees of pheromones and supports on corresponding path in every level，and gets node’s preference function values，then finds user interest model by this value.Experiment results show that the method is valid and practicable.

摘要： 为了得到更多的用户兴趣信息，提出了一种新的高效的用户兴趣模式获取方法，面向层次结构的信息网站，提供动态的多层次用户兴趣模式。该方法根据网站和用户兴趣所具有的层次性特征，利用蚁群算法，将用户对网站结点的一次浏览过程，对应于蚂蚁的一次觅食活动周期，从各个层次求出相应路径的信息素浓度和支持度，从而得到用户对该结点的偏好函数值，再依据此值求得用户兴趣模式。实验表明了该方法的有效性。