Machine Learning-based Multi-Module System for Student Depression Prediction and Behavioural Analysis

More students are being affected, and they are finding themselves depressed, stressed, anxious, having sleeping issues as well as unhealthy digital behaviour patterns that directly impact their study life and overall wellbeing. Manual intervention Traditional screening processes are known to be time consuming, inconsistent and not always linear. This research proposes an intelligent mental health decision support system, which is inclusive of Machine Learning (ML), behavioural analytics, and AI-assisted recommendation strategies in the early detection of depression risk and digital wellbeing conditions among students. The proposed system is made up of two large prediction modules. The first module approximates the risk of depression with the help of personal, academic, lifestyle, and family related variables age, gender, city, degree, CGPA, academic stress, financial stress, sleep, diet, work-study hours, suicidal thoughts, and family history of mental illness. The second module measures digital wellbeing by behavioural attributes such as average time spent on the screen, frequency of switching between apps, amount of time spent asleep, the number of notifications, time spent on social media, focus score, mood score and the level of anxiety. The system also has a rule-based overriding process, which is used to detect emergency cases particularly in cases where there are suicidal thoughts or the presence of powerful stressful conditions. A recommendation engine based on AI will provide personalized mental health recommendations, lifestyle recommendations, information on support, and motivational content depending on the level of predicted risk. There is also storage of historical prediction, behavioural timelines and risk trend visualization, which provides long term monitoring and analysis. The suggested system is a scalable model offering an efficient and easy to use model of early mental health evaluation and constant behavioural monitoring.