标签关联与学习模型相结合的用户兴趣变化识别算法  

User Interest Change Recognition Algorithm Based on Combining Tag Association and Learning Model

作　　者：白杨[1] BAI Yang(School of Information Engineering,Eastern Liaoning University,Dandong 118003,China)

机构地区：[1]辽东学院信息工程学院,辽宁丹东118003

出　　处：《辽东学院学报（自然科学版）》2021年第3期180-186,共7页Journal of Eastern Liaoning University:Natural Science Edition

基　　金：辽宁省教育科学“十三五”规划课题(JG20DB166);教育部产学合作协同育人项目(202002273008)。

摘　　要：通过分析用户的UGC内容获得用户的兴趣是目前一个研究热点,识别社交网络的用户兴趣变化对于提供高质量的网络个性化服务具有重要意义。对于UGC贫乏的“冷启动”用户,常用的LDA主题模型获取用户兴趣并不有效,而能有效识别用户兴趣的变化更是一个难点问题。针对这一问题,采用LDA获取网络主题,构建主题-标签模型,在此基础上采用基于标签关联的空间向量相似度计算方法,构建用户-主题模型,并结合学习模型,提出用户兴趣变化识别算法。在实验中,将所提出的算法应用于从微博网站所采集的大规模数据集,较为准确地获得了网络的主题、网络主题的核心用户及其在兴趣主题上的概率分布,有效地解决了模型中用户兴趣变化的识别问题。It is a hot research topic to obtain user interest by analyzing user’s UGC content.Identifying user interest changes in social networks is of great significance for providing high-quality personalized network services.For the‘cold-start’users with poor UGC,LDA topic model is not effective to obtain their interest,and it is more difficult to identify the interest changes of them.To solve this problem,LDA is firstly used to obtain network topics,and by which the topic-tag model is constructed.Furthermore,the spatial vector similarity calculation method based on tag association is used to construct the user-topic model.Combined with the learning model,a user interest change recognition algorithm is proposed.The proposed algorithm is applied to the large-scale data set collected by the microblog website,and the network topic,the core users of the network topic and their probability distribution on the interest topic are obtained more accurately,thus the identification problem of user interest change in the model is solved.

关 键 词：主题模型 用户兴趣 空间向量 学习模型 标签关联 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]