基于知识图谱的社交媒体内容可信审查研究  

Study on Credibility Review of Social Media Content Based on Knowledge Graph

作　　者：熊建英[1] 涂敏[1] XIONG Jianying;TU Min

机构地区：[1]江西警察学院安全技术系,江西南昌330003

出　　处：《图书情报导刊》2020年第10期35-40,共6页Journal of Library and Information Science

基　　金：教育部人文社科青年基金项目“基于信任与内容风险感知的社交媒体传播治理研究”(项目编号:20YJC860032), 江西省高校人文社会科学规划项目“基于用户信任的社交网谣言抑制研究”(项目编号:GL18121)。

摘　　要：社交网已经成为人们获取信息的重要渠道,其无障碍的互动模式也使得不实信息快速传播,给社会造成了负面影响。构建了基于知识图谱的网络大数据信息组织方式,利用知识一致性对社交媒体内容的可信性进行了自动审查。首先,从平台管理和用户角度制定了知识管理需求,定义了内容实体类型包括内容主题、内容主体、来源方、传播方及各实体之间关系和属性,构建了社交内容的本体论模型, 其次,基于本体概念模型构建了社交媒体内容知识图谱管理架构以实现网络开放知识集成, 最后,以案例论证如何利用内容知识图谱实现对不可信内容的审查。结果表明:以知识图谱方式组织社交媒体内容,有利于提升对文本的语义理解,提高审查的可解释性。Social network has become an important channel for people to obtain information.The free interactive com⁃munication mode also makes false information spread rapidly,which causes serious negative impact.A knowledge graph based on network big data information organization is built,and consistency knowledge is used to automatically review the credibility of social content.This paper formulates the knowledge management requirements from the perspective of plat⁃form management and ordinary users,defines the content entity types including content subject,content subject,source,disseminator and the relationship and attribute between each entity,builds an ontology model of social content,and then based on which,builds a management framework of knowledge graph of social media content to integrate the existing open knowledge,and finally demonstrates how to construct the content knowledge graph to realize the review of untrusted con⁃tent in the mode of case.Research examples show that it is conducive to improving the semantic understanding of the text and improving the interpretability of censorship.

关 键 词：社交媒体 知识图谱 内容可信 内容安全 

分 类 号：G206.2[文化科学—传播学]