基于公开信息的微博用户可信性评价研究  

Research on the Evaluation of Micro-Blog Users' Credibility Based on Public Data

作　　者：赵丽华[1] 杨勇[1,2] 闻西 李卫康[3] Zhao Lihua;Yang Yong;Wen Xi;Li Weikang(School of Management and Economics, Tianjin University of Technology and Education, Tianjin 300222 , China;College of Management and Economics, Tianjin University, Tianjin 300072 , China;Business School of Human University, Changsha 410082 , China)

机构地区：[1]天津职业技术师范大学经济与管理学院,天津300222 [2]天津大学管理与经济学部,天津300072 [3]湖南大学工商管理学院,长沙410082

出　　处：《天津大学学报（社会科学版）》2018年第3期217-223,共7页Journal of Tianjin University：Social Sciences

基　　金：教育部人文社会科学青年基金资助项目(15YJC790131);天津市科技特派员基金资助项目(15JCTPJC61400);国家自然科学基金资助项目(71171076);天津哲学社会科学规划基金资助项目(TJGL15-049);天津市哲学社会科学基金资助项目(TJSR13-012)

摘　　要：基于微博用户公开的个人资料,应用数据挖掘方法,构建了非认证微博用户可信度评价模型。该模型从权威性和真实性两个角度对用户公开的个人资料数据进行划分,对原始数据进行不同的归约和转换处理,从统计角度,考察了各个数据信息项与用户可信性之间的关系,并以此为依据构建了微博用户可信度评价特征集合。实验结果表明,该模型可有效用于度量微博用户的可信性。Based on the micro-blog users ’ individual data, this paper constructs an evaluation model of the non-certified micro-blog users ’ credibility using the data mining method. This model divides the users ’ public individual data into two aspects ： authority and authenticity and processes the original data through reduction or conversion. This model also analyses the relationship between the individual information and the users＇ credibility in a statistical perspective and builds a micro-blog users＇ credibility evaluation feature set based on these. The experimental results show that this model can effectively evaluate the non-certified users＇ credibility.

关 键 词：可信度 特征抽取 微博 数据挖掘 

分 类 号：F125[经济管理—世界经济]