中文摘要：

基于标签进行个性化推荐是目前的一个研究热点，不同的推荐算法对标签进行了不同的处理。用户使用的标签之间存在着某种内在联系，由此可构建用户标签网络。根据这一启示，本文提出了一种基于用户标签网络的个性化推荐算法。首先，将用户标签网络视为用户兴趣模型雏形，利用社会网络分析方法计算标签权重，并以加权标签集的形式表示用户兴趣模型，最后将标签权重转化为资源与用户兴趣的相似度，进而实现个性化推荐。实验表明，本方法能较为准确地揭示用户的兴趣，产生的推荐资源与用户兴趣匹配程度较高。

英文摘要：

Research on personalized recommendation based on tag has always been a hot topic. Different algorithms calculate tags differently. There are some essential associations between tags generated by a user, based on which the user tag network can be constructed consequently. Inspired by this idea, this paper designs a personalized recommendation algorithm. The user tag network is viewed as the archetype of user＇s local interest model. Then, the weight of tag is generated by applying the social network analysis. In this paper, the user interest model is represented by the weighted tags. The weight of tag is finally transformed into the similarity between items and user, which results in the implementation of personalized recommendation. The experiment shows that this approach can accurately describe user＇s interest and generate well-matched items for user.