[1Win4] Poster session 1

[1Win4-34] Feature Extraction of Tweets from Groups with Mental Health Issues Using Large Language Models

〇Honoka Takabatake¹, Katsuhito Sudoh¹, Yoshinobu Kano² (1.Nara Women's University, 2.Shizuoka University)

Keywords:Large Language Models

This study aims to extract characteristic tweet expressions by those suffering from mental health issues using
a large language model. We use a large-scale dataset of tweets from individuals with and without mental health
issues for this purpose. We first extract keywords indicating mental distress from a large portion of the dataset
and then examine those keywords on the rest. The results suggest that using a large language model can extract
characteristic expressions effectively.