一种基于用户行为的兴趣度模型

摘要

Personalized recommendation is a widely applied technology in e-commerce. Since the existing user models can not make correct recommendation from the consumer interest, the paper presents a personal interest degree model based on consumer behavior, which analyzes consumer behavior mode, and combines with the contents the user has browsed, to discover the contents consumer interested. On this basis, it uses the expectation-maximization algorithm to create user clustering and assign the user to the corresponding cluster for construction of the interest degree model, which can be further used for implementation of personalized recommendation to users. Experimental results show that the model can identify users personalization interests better, since it can improve the recommendation accuracy and customer satisfaction.%个性化推荐技术在电子商务系统中得到了广泛应用.针对现有的用户模型不能根据用户自身兴趣实现推荐的问题,提出了一种基于用户行为的兴趣度模型,分析用户的行为模式,结合用户的浏览内容,发现用户兴趣.在此基础上采用期望最大化算法实现用户聚类,将用户划分到对应的簇,创建用户的兴趣度模型,从而向用户进行个性化推荐.实验对比结果表明,该模型能更好地发现用户当前的购买兴趣,从而进一步提高个性化推荐精度和用户满意度.