A Comparative Study of Personality Assessment Based on Different AI Models

• In Event: Virtual Poster Session
  In Virtual iPresentation Only: Division D - Section 1: Educational Measurement, Psychometrics, and Assessment Virtual Poster Session

Abstract

Accurate personality assessments are essential for personalized learning plans and educational equity. Traditional methods are reliable but often time-consuming and biased. This study explores the use of DeBERTa and ChatGPT-4 for personality assessment with data from Zhihu (a Chinese question and answer platform). We identified seven new dimensions—Sophistication, Kindness, Depth of Thought, Self-Control, Eagerness to Learn, Desire for Success, and Pursuit of Happiness—refined using BERTopic and the Delphi method. DeBERTa excels in deep semantic analysis (e.g., Sophistication F1: 0.768, Depth of Thought F1: 0.832), while ChatGPT-4 better captures emotional nuances (e.g., Self-Control F1: 0.670, Pursuit of Happiness F1: 0.821). This study further suggested that combining these models can improve personality assessment accuracy.

Authors

• Tingzhang Wu, Beijing University of Chinese Medicine

• Lin Wu, Beijing University of Chinese Medicine

• Yixing Liu, Beijing University of Chinese Medicine