沈旺 代旺 高雪倩

摘 要：[目的/意义]提出基于多重图的社交网络用户可信度评价方法，为网络信息生态治理提供有效依据，为有害次生舆情管理提供理论基础。[方法/过程]构建了网络欺凌和隐私泄露特征抽取指标体系;利用多重图表示社交网络用户动态信息网络;根据提取到的网络欺凌和隐私泄露信息进行用户信息行为分析;利用德尔菲法确定网络欺凌行为与隐私泄露行为概率的权重，综合分析社交网络用户可信度。[结果/结论]实验结果表明，基于多重图的用户可信度评价结果与基于调查问卷的用户可信度评价结果基本一致，说明利用多重图的社交网络用户可信度评价方法具有一定的可行性。

关键词：社交网络;多重图;网络欺凌;隐私泄露;用户可信度

Abstract：[Purpose/Significance]A credibility evaluation method for social network users based on multigraph was proposed，aiming to provide an effective method for ecological governance of network information and a theoretical basis for the management of network secondary public opinion with negative effects.[Method/Process]The paper constructed the index system of feature extraction of cyberbullying and privacy disclosure，and the dynamic information network of social network users was expressed by using multigraph.The extracted information of cyberbullying and privacy disclosure was used to evaluate information behavior of network users.The probability weight of cyberbullying and privacy disclosure was determined by Delphi method，and the credibility of social network users was comprehensively analyzed.[Result/Conclusion] The experimental results showed that the result of the users credibility evaluation based on multigraph was basically the same as the result obtained by the questionnaire，indicating that it is feasible to use multigraph to evaluate the credibility of social network users.

Key words：social network;multigraph;cyberbullying;privacy disclosure;credibility of users

随着移动互聯网的发展，微博和知乎等社交网络平台已成为我国大多数人获取信息和发表言论的重要平台，成为网络舆论及意识形态争论的高地。目前，由某热点事件衍生次生舆情的情况十分常见，此类舆情往往会产生“倒逼”性的压力，致使当事主体难以应对[1]。

从狭义上来讲，所谓次生舆情是当原生的舆情为负面事件，亦或是可能引起负面次生舆情时所产生的关联舆情情况[2-3]。网络负面舆情常常会衍生出包含网络欺凌和隐私泄露的次生舆情，如在新冠肺炎疫情期间，部分社交媒体用户对作出其不认同行为或发布其不认同信息的用户进行谩骂甚至泄露其他用户的隐私。……