中文摘要：

随着移动互联网的发展，越来越多的用户信息获取过程通过移动终端完成。但当前个性化推荐系统对用户情境的感知能力不足，缺乏为用户提供符合当前情境的个性化信息推荐服务。为此，本文提出了基于贝叶斯方法的情境化用户资源类别偏好学习以及融合该类别偏好的协同过滤个性化信息推荐。运用贝叶斯方法学习用户在不同情境下对各资源类别的偏好，然后将该类别偏好与传统协同过滤推荐算法相结合，生成符合用户当前情境的个性化信息推荐。实验表明本文提出的改进算法可以提高推荐的准确率。

英文摘要：

With the development of Mobile lntemet, more and more people acquire information from Web via mobile terminals. Lacking of context awareness,current personalized recommendation systems can ＇ t offer adaptive services. This paper presents the learning of contextual resource category preferences Of users based on Bayesian method. And we combine the preferences with traditional collaborative filtering. With Bayesian approach we explore users＇ preferences in resource categories under different conditions and the combination with collabora- tive filtering enables personalized information service to develop its context-aware ability. The experiment shows that the proposed algo- rithm can improve the accuracy of recommendation.