A Study on User Classification based on The Content Posted to SNS using Self-Organizing Map

Abstract

In recent years, the number of users of matching apps(free-market, marriage hunting, etc.) has increased. Many matching apps match users based on the profile information entered by the user for predetermined items. At this time, the profile information is only a self-report and does not necessarily accurately represent the objective characteristics of the user, so the validity of the matching result may be doubtful. In this research, we try to vectorize each user using a method called Word2Vec based on the text data posted by the user on SNS, and visualize the characteristics of each user with a Self-Organizing Map(SOM).