Artificial Intelligence Sensor for “Big Five” Personality Analysis

Behavior measurement has played a significant role in research within the field of personality psychology, providing objective and quantifiable data that helps us gain deeper insights into individuals' behavioral characteristics and psychological states. However, assessing personality through gait analysis remains a significant challenge. In this study, we developed a dynamic monitoring system for gait parameter analysis. The inertial measurement unit embedded in the system consists of a three-axis accelerometer and a three-axis rotation-rate gyroscope. The sensor was placed on the participants' ankle to investigate the relationship between walking gait and personality. Foot motion data, including acceleration and angular velocity, was collected during the natural walking process. Additionally, with the 60-item Big Five Inventory-2 (BFI-2), participants' personality traits were assessed. Classifiers with different learning algorithms were used to examine the relationship between foot motion features and various personality traits. The results showed that the Random Forest (RF) algorithm performed best in classifying the five dimensions of personality, and the accuracy for the agreeableness dimension reached an impressive 89.66%. This research provides a new tool for objective and in-depth understanding of personality through artificial intelligence sensors.