Methods for Extracting Opinion Trends to Promote the Experience in Determining the Authenticity of SNS Articles

There are various articles on social media that are factually incorrect mixed in with factual and informative articles. Therefore, it is important to identify fake articles when browsing the numerous posts on social media. Although unskilled people need to gain a lot of experiences in examining the authenticity of articles in order to obtain the ability to distinguish it, it is often difficult for novices to stack up enough meaningful experience to do so. We therefore aim to develop methods that promote the experience of examining the authenticity of articles by presenting users with articles whose authenticity should be questioned along with other people's opinions on those articles. In this paper, we mainly describe methods for collecting candidates of social media articles that may include incorrect descriptions and others' opinions, along with methods for classifying the collected articles in accordance with opinion trends. Finally, we describe an experiment utilizing the proposed methods for classifying opinion trends and discuss their effectiveness of our methods based on its results.