Too Neurotic, Not Too Friendly: Structured Personality Classification on Textual Data

Abstract

Personality plays a fundamental role in human interaction. With the increasing amount of online user-generated content, automatic detection of a person's personality based on the text she produces is an important step to labeling and analyzing human behavior at a large scale. To date, most approaches to personality classification have modeled each personality trait in isolation (e.g., independent binary classification). In this paper, we instead model the dependencies between different personality traits using conditional random fields. Our study finds a correlation between Agreeableness and Emotional Stability traits that can improve Agreeableness classification. However, we also find that accuracy on other traits can degrade with this approach, due in part to the overall problem difficulty.