[3A5-GS-6] Language media processing

[3A5-GS-6-03] Content Analysis of Comments Posted on Internet TV News Programs

〇Rin Atsuji¹, Yuki Ogawa¹, Fumiaki Taka³, Masanori Takano², Soichiro Morisita² (1. College of Information Science and Engineering,Ritsumeikan University, 2. CyberAgent, Inc., 3. Faculty of Sociology, Toyo University)

Keywords:ToxicScore, internet TV, news

Live video viewing via Internet TV is becoming a popular way to watch news. There, a chat function may exist where viewers can freely post comments in real time. Such a function is expected to have positive aspects, such as viewers posting their opinions and impressions to enliven the program and gain knowledge through communication. However, there is also concern that the flow and atmosphere of comments posted in the chat section may encourage excessive and excessive comments. If the flow of chat comments posted in the future can be predicted, it will be possible to identify in advance situations in which chat sections become rough and to consider countermeasures such as warning people to be careful. As a basic analysis for this purpose, this study analyzes the contents of chats posted on ABEMA news programs, and analyzes the differences depending on the news format and the content of the news in terms of the positives and negatives of the posted text, sentiment, and Toxic Score.