Development of a Bot using Sentiment Dictionary and Machine Learning for Proactive Detection of Slanderous Japanese Posts on Social Media
M. Fahim Ferdous Khan, Ryo Inoue, Ken Sakamura · 2021 IEEE 10th Global Conference on Consumer Electronics (GCCE) · 2021
Slanderous posts and other forms of abusive online behavior have become prevalent in recent times. We emphasize that detecting and deleting slanderous messages after they have been posted on social media is not enough as the damage may have already been inflicted on the victims in the meantime. Of course, there are deliberate slanderers on the Internet, but many people run the risk of becoming unintentional slanderers for having no or inadequate knowledge about what may constitute an abusive online behavior. For such uninitiated users, a social media tool for checking whether a to-be-posted message can be deemed potentially slanderous would be useful. Therefore, in this paper, we present the design and implementation of a social media bot for proactively detecting slanderous post written in Japanese. Our bot combines the use of sentiment dictionaries and state-of-the-art machine learning techniques. A preliminary evaluation of the bot showed promising results.