Identifying Predictive Factors of Mood Swings and Mental Health

Abstract

Mood swing is one of the common symptom related to mental health. It can be influenced by biological, psychological, and environmental factors. However, the process of identifying health is challenging and varies between individuals. Therefore, machine learning was employed to analyze the contributing factors to different level of mood swing. Gender, days indoors, growing stress, habit changes, coping struggles, work interest, and social weakness are among the important features found. It was reported that K-Nearest Neighbor performed the best across all metrics (accuracy=95.77%, precision=95.89%, recall=95.77%), suggesting that a distance-based approach is well-suited for this dataset followed by Support Vector Machine, Decision Trees and Random Forest. Consequently, early detection and targeted interventions would significantly improve the quality of life for affected individuals.