Affective Disorder Detection in College Students with Weak-Privacy Rhythmical Behavioral Data from Smartphones

As affective disorders become increasingly prevalent among college students, the management of these disorders on university campuses has become a societal concern. The stressors contributing to affective disorders faced by college students are diverse and often lack obvious symptoms or signs, highlighting the critical importance of addressing these issues in a timely manner. Previous research has established a correlation between smartphone sensor data and the emotional state of college students. Considering the importance of privacy protection for students, this study attempts to assess the affective disorder status of college students based on their weak-privacy data collected on campus. We developed a smartphone application to collect weak-privacy behavioral data from smartphone sensors, extract daily behavioral features, and applied various machine learning methods to evaluate the emotional status of students. This research provides a new perspective and approach for research and intervention in the field of affective disorders within campus environments. By leveraging machine learning techniques and behavioral data, this study helps to develop proactive strategies for identifying and supporting college students at risk of affective disorders, ultimately promoting mental health on university campuses.