Kurniati Rahayuni, Desi Fatkhi Azizah, Yukta Terate Syura, Siti Nurrochmah
This study explores the use of machine learning algorithms to classify athletes' psychological states based on the Athletic Coping Skills Inventory-28 (ACSI-28) scores. Psychological states significantly impact athletic performance, making accurate monitoring and assessment essential. Four machine learning algorithms were employed: Decision Tree, Random Forest, Support Vector Machine (SVM), and Naïve Bayes, using both ACSI-28 and demographic data. Results indicated that Random Forest performed best, achieving the highest accuracy (0.807692) and F1-Score (0.81491) in a 70:30 data split after parameter tuning. SVM demonstrated stable accuracy (0.779923) and better time efficiency in the 50:50 split. The Decision Tree showed moderate accuracy (0.779923) but excelled in interpretability. Although Naïve Bayes was time-efficient (0.036661 seconds), it yielded the lowest accuracy (0.679487). Parameter tuning improved performance, with Random Forest leading in accuracy, suggesting machine learning's potential for enhancing personalized mental training programs. ©2025 IEEE.
Department of Sports Coaching Education, Universitas Negeri Malang, Malang, Indonesia; Department of Electrical Engineering and Informatics, Universitas Negeri Malang, Malang, Indonesia; Department of Physical Sports Education, Universitas Negeri Malang, Malang, Indonesia
