A machine learning-based approach to predict university students' depression pattern and mental healthcare assistance scheme using Android application

Depression is particularly common among university students in developing countries like Bangladesh. University students may face challenges with their studies, relationships, drugs, and family issues, all of which are major or minor contributors to depression. This research study focuses on gaining useful insights into why university students in Bangladesh suffer from depression and predicting depression in university undergraduates for the purpose of referral to a psychiatric facility. A Google survey form was used to gather data for this study. After training and testing the dataset with five algorithms, the best methods for predicting depression among Bangladeshi undergraduate students were discovered. A comparison of various prediction algorithms such as logistic regression, KNN, SVM, random forest, decision tree, including accuracy, precision, recall, error rate, f-measure, mean absolute percentage error for analysis was done. We also designed and developed an Android mental healthcare mobile application to provide mental support to university students.