A Novel Method based on Large Language Model for MBTI Classification

Personality analysis has a wide and important application in psychology, helping us explain various psychological phenomena and the developmental laws of personality. Accurately identifying personality will contribute to research in fields such as cognitive science, public opinion analysis, and cybersecurity. The most widely used models in the field of personality recognition are the Big Five personality model and MBTI model in trait genre. With the rise of social media and Large Language Model(LLM), massive corpora and deep learning models have been used for personality analysis and have achieved good results. In view of this, this article introduces a novel method based on transformer-based pre-trained language model named mDeBERTa and uses the MBTI-500 and MBTI-1 datasets as examples to demonstrate the advantages of this model over traditional SVM models and BERT models. At the same time, we explore the possibility of using large models and multimodal data for more accurate personality analysis.