Topic Model Analysis of Japanese Microblog Profile Text

Abstract

Analyzing the data in SNS has played a very important role in marketing and other areas. In these analyses, it is expected that the content of posts and user profiles can be used together to further increase accuracy. For this reason, this paper classifies profiles into topic categories and clarifies what kind of content the profiles consist of. Topic model analysis revealed some characteristics of the profile texts, such as some profile texts describing male idols and some profile texts telling the persona of the twitter user.