Analysis of Articles that Correct Other Posts on Social Media Aimed at Promoting the Experience in Examining Fakes

Social media is increasingly being used as a tool to gather a wide variety of information. However, there are fake articles on social networking services mixed in with useful posts. It is desirable for users to use social networking services while determining the truth or falsity of articles. However, such judgement is difficult for inexperienced users since the skills to determine the authenticity of articles should be obtained by a stacking of experiences. In this research, we aim to develop methods for gaining experience with examining fake articles by suggesting noteworthy articles on the basis of an analysis of others’ responses to the articles. This paper describes methods for extracting articles that correct other posts on the basis of the characteristics of people’s responses to articles on social networking services and for extracting candidates for fake articles by analyzing such articles. Finally, we describe an experiment using a prototype system and discuss the effectiveness of our system as based on its results.