An NLP model that understands stories should be able to understand the characters in them. To support the development of neural models for this purpose, we construct a benchmark, Story2Personality. The task is to predict a movie character's MBTI or Big 5 personality types based on the narratives of the character. Experiments show that our task is challenging for the existing text classification models, as none is able to largely outperform random guesses. We further proposed a multi-view model for personality prediction using both verbal and non-verbal descriptions, which gives improvement compared to using only verbal descriptions. The uniqueness and challenges in our dataset call for the development of narrative comprehension techniques from the perspective of understanding characters.
@article{arxiv.2210.10994,
title = {MBTI Personality Prediction for Fictional Characters Using Movie Scripts},
author = {Yisi Sang and Xiangyang Mou and Mo Yu and Dakuo Wang and Jing Li and Jeffrey Stanton},
journal= {arXiv preprint arXiv:2210.10994},
year = {2022}
}