用户隐藏属性推断研究综述  

Survey on Latent Attribute Inference of Social Media Users

作　　者：宋巍[1] 谢兴波 刘丽珍[1] 王函石[1] 

机构地区：[1]首都师范大学信息工程学院,北京100048

出　　处：《小型微型计算机系统》2017年第2期217-226,共10页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(61402304;61303105)资助;北京市自然科学基金项目(4154065)资助;教育部人文社会科学规划项目(14YJAZH046)资助;北京市教委科研支持项目(KM201610028015)资助

摘　　要：社交媒体的广泛使用,积累了大量的用户数据,为深度挖掘和分析海量异构社交网络带来了巨大的机遇,用户隐藏属性推断应运而生.用户隐藏属性推断,旨在自动预测用户的未知属性与潜在特质.总结了基于社交媒体的用户隐藏属性推断的最新进展.首先介绍了用户隐藏属性推断相关研究;进而将用户隐藏属性推断归纳为三项主要任务,即数据采集与筛选、特征设计和推断方法,并对它们进行了细致的介绍和分析;介绍了用户隐藏属性推断的应用;最后总结并分析了未来的研究方向.重在对用户隐藏属性推断研究的主流方法和前沿进展进行概括、比较和分析.The wide use of social media has accumulated a lot of user data, which provides great opportunities for the deep mining and analysis of huge amounts of heterogeneous social network. User latent attribute inference hidden attribute inference has emerged at the historic moment and has become a new important research topic. User latent attribute inference alms at predicting users＇ unknown at- tribute and latent trait automatically. In this paper, we summarize the recent progress of social media based user latent attribute infer- ence. Firstly, the paper introduces the research situation of latent attribute inference, which mainly presents survey on user latent attrib- ute inference in different areas. Then, three important tasks of user latent attribute inference are summarized and analyzed in detail, in- cluding data collection and selection, feature design,inference method. The applications of latent attribute inference are concluded. Finally, we carry out a preliminary prospect on the research trends in future. This paper aims to take a deep insight into the mainstream methods and recent progress in this field,making detailed comparison and analysis.

关 键 词：用户隐藏属性 用户性别推断 用户偏好建模 社交媒体 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]