<![CDATA[This study proposes an anxiety prediction model based on smartwatch activity data integrated with a self-reported affect scale in adolescents. Anxiety among adolescents is a growing public health concern, often underdetected due to subjective assessment and limited continuous monitoring. To address this gap, this research combines objective physiological and behavioral indicators collected from smartwatches, including heart rate variability, sleep duration, physical activity intensity, and daily movement patterns, with subjective emotional states measured through a validated affect scale. Data were collected longitudinally from adolescent participants over several weeks to capture temporal variations in activity and mood. Machine learning techniques were applied to develop and evaluate predictive models capable of identifying anxiety levels with high accuracy. Model performance was assessed using standard metrics such as accuracy, precision, recall, and F1-score. The results demonstrate that the integration of wearable sensor data with self-reported affect significantly improves anxiety prediction compared to single-source data models. The proposed model offers a scalable, non-invasive, and real-time approach for early anxiety detection, supporting timely intervention and personalized mental health monitoring for adolescents. This study contributes to the development of human-centered digital health technologies and highlights the potential of wearable-based analytics in preventive mental healthcare systems for future applications.]]>
