Identifying Opponent's Neuroticism Based on Behavior in Wargame

Area of Science:

• Psychology
• Machine Learning
• Behavioral Science

Background:

• Traditional neuroticism assessments rely on self-report questionnaires, which are limited in confrontational settings and prone to bias.
• There is a need for objective and practical methods to assess neuroticism, especially in competitive environments.

Purpose of the Study:

• To develop and validate a machine learning-based approach for predicting neuroticism using behavioral data.
• To overcome the limitations of subjective self-report measures in assessing neuroticism.

Main Methods:

• Analyzed behavioral records from 167 participants on the MiaoSuan Wargame platform.
• Identified key behavioral features associated with neuroticism and developed predictive models.
• Utilized the 8-item neuroticism subscale of the Big Five Inventory for neuroticism assessment.

Main Results:

• The best-performing model, LinearSVR, effectively inferred neuroticism levels from behavioral data.
• Achieved a correlation of 0.606 between predicted and self-reported neuroticism scores.
• Demonstrated test-retest reliability of 0.516, indicating model stability.

Conclusions:

• Behavioral data can be effectively used to predict neuroticism, offering a viable alternative to self-report measures.
• The developed method has practical applications in psychological assessment within competitive environments.
• Future research should focus on refining feature selection and expanding application scenarios for behavior-based assessment.