EARLY DETECTION OF ACADEMIC DEPRESSION USING SMARTPHONE-BASED MACHINE LEARNING MODELS

Mental health in developing countries is a common and complex problem. The problem continues to increase and is closely related to low self-confidence, negative interpersonal relationships, and academic depression. This can affect students' ability to complete academic assignments on a university scale. An AI-based early detection application can potentially improve mental health services related to treatment access. This system can help identify users who may be depressed based on the language used, especially for those who are reluctant to seek professional solutions due to the negative stigma of mental health. This study uses a qualitative descriptive method involving observation, in-depth analysis of group conversations, and early detection of academic depression by identifying conversation patterns between students and counselors as the basis for developing a smartphone-based application. This study produced a dataset of 395 depression-level data entries used as training data to develop a machine-learning model. A prototype of an academic depression detection application has been developed as a mobile-based application.