Multivariate Optimization of Proactive Attitude Toward Learning Evaluation System Using Biometric Data

This study explored the relationship between evaluations of a proactive attitude toward learning-Japan’s learning assessment criteria-and biometric data such as blink count, head sway, and electroencephalogram (EEG) signals. We previously developed a system that quantitatively evaluates self-study using biometric data and improved it by optimizing system threshold adjustments to accommodate different school environments and student groups. This study examined the application of EEG evaluation to different learning environments. Furthermore, the system evaluation thresholds were parameterized and multivariate optimization was used to enhance accuracy and reduce the user burden. Blink count-based evaluations during calculation tasks demonstrated potential for producing learning attitude assessments comparable to teacher judgments. Potential overfitting in noncalculation task evaluations and EEG assessments indicate the need for further verification with a participant sample larger than that considered in this study.