摘要： 随着移动计算技术的发展，人们可以在移动环境中方便地在线获取阅读资源，但如何在海量资源中检索出符合用户兴趣的内容，成为亟需解决的问题。为此，提出一种面向移动阅读平台的资源推荐算法。根据用户的知识结构和用户之间的交互记录进行建模，计算用户相似度以获取相似用户，利用最近邻集合结合协同过滤算法进行资源推荐。在系统平台上进行测试，该算法的绝对误差平均值为0.636，低于同类推荐系统的平均水平，表明推荐算法是有效的。

关键词: 阅读平台, 用户相似度, 知识结构, 交互记录, 协同过滤, 资源推荐

Abstract: The development of mobile computing technology makes people able to get online reading resources in a mobile environment. It is becoming a problem how to retrieve contents matching users’ interest from massive resources, therefore, a resources recommendation algorithm oriented mobile reading platform is proposed. The algorithm applies the feature of users’ knowledge structure and social intercommunication records into the calculation of similarity between users to get the nearest-neighbor set of the collaborative filtering method. Test result on the system platform shows that absolute error average value of the proposed algorithm is 0.636, it is lower than average level of recommendation system, the recommendation algorithm is effective.

Key words: reading platform, user similarity, knowledge structure, interaction records, collaborative filtering, resources recommendation

中图分类号: 

• TP311