“What is your MBTI?”: Predicting the personality types using hierarchical attention and graph learning

The Myers—Briggs type indicator (MBTI) is a widely-used personality test that classifies user personality types into one of the 16 categories. This study first investigates whether people with identical MBTI types show similar linguistic features across different social media (i.e., Kaggle, Twitter, and Reddit). Based on the lessons learned from analysis, we propose a model that can predict a user's MBTI type from their social media text data. To learn the linguistic characteristics of each personality type, we introduce a personality vocabulary graph that represents the relationship between used words and each personality type. We utilize hierarchical attention in the proposed model to highlight important words and sentences that reveal the user's personality. The results demonstrate that the proposed model outperforms baseline models across the different social media, implying that the proposed model is effective in user-level personality prediction and generally applicable in various social media. This study can provide important implications on user profiling, which is essential in targeted marketing, recommender systems, and political campaigns.