[A-11] Thought and Language

10:30 AM - 10:45 AM JST(1:30 AM - 1:45 AM UTC)Candidate for Academic Incentive Award

[A-11-01] Analysis of Linguistic Expression Patterns in SNS Posts for MBTI Tendency Prediction

◎Ayaka Sato¹, Takahiro Baba¹ (1. Kurume Institute of Technology)

Keywords:

MBTI,Social Media Analysis,Text Mining,Personality Prediction

Recently, personality prediction using social media posts has attracted considerable attention. In this study, we analyzed linguistic expression patterns in SNS posts associated with MBTI personality types. A total of 1,600 posts were collected from X (formerly Twitter), with 100 posts for each of the 16 MBTI types. Morphological analysis and TF-IDF were applied to extract characteristic words for each type. The results revealed distinctive expressions such as “thinking,” “suffering,” and “talking deeply” for INFP; “nature,” “adventure,” and “curiosity” for ISFP; and “ideal,” “study,” and “leadership” for ENTJ. Furthermore, classification experiments using TF-IDF features achieved a Macro F1-score of 0.228 for 16-type classification. For the four MBTI dimensions, the highest accuracy was obtained for the Thinking/Feeling (T/F) dimension at 62.9%. These findings suggest that MBTI-related tendencies are partially reflected in linguistic expressions on social media.