基于标签与关系网络的用户聚类推荐研究

发布时间：2019-03-18 21:50

【摘要】：【目的】利用用户标签及关系网络,为用户推荐潜在的相似用户。【方法】通过探究社会化标注系统中标签、关系网络所表征的用户长短期兴趣特征,综合用户标签及关注关系,利用多维尺度法构建用户聚类模型,根据用户聚类结果进行相似用户推荐,并以"微博"为例对模型进行实证。【结果】实验结果表明,基于标签和关系网络的用户聚类模型能够有效地结合用户长短期兴趣特征,挖掘潜在相似用户,聚类及推荐效果较好。【局限】样本数据集具有局限性,不能完全涵盖用户兴趣领域,仅从一个领域验证了模型的准确性与有效性。【结论】通过对用户标签及关系网络挖掘用户长短期兴趣,构建的基于用户静态标签与动态关系网络的用户推荐模型,对个性化用户推荐效果有较好的提升。
[Abstract]:[objective] to recommend potential similar users to users by using user tags and relational networks. [methods] by exploring the long-term and short-term interest characteristics of users represented by relational networks in socialized labeling system, According to the user label and the relation of concern, the user clustering model is constructed by using multi-dimension scale method. According to the result of user clustering, similar user recommendation is made, and "Weibo" is taken as an example to demonstrate the model. [results] the experimental results show that: The user clustering model based on label and relational network can effectively combine the long-and short-term interest characteristics of users, mining potential similar users, clustering and recommendation effect is better. [limited] sample data set has limitations. The user interest domain can not be completely covered, and the accuracy and validity of the model can be verified from only one domain. [conclusion] by mining the long-term and short-term interest of users through user tags and relational networks, The user recommendation model based on user static label and dynamic relation network can improve the effect of personalized user recommendation.
【作者单位】：华中师范大学信息管理学院;
【基金】：国家社会科学基金项目“大众分类中标签间语义关系挖掘研究”(项目编号:12BTQ038)的研究成果之一
【分类号】：TP391.3

【相似文献】