Depression is a prevalent mental health issue among university students, negatively affecting academic performance, social interactions, and overall quality of life. Traditional assessment methods are often time-consuming and limited by subjective responses. This study aims to predict university students' depression status using various machine learning algorithms. The dependent variable is depression status (Yes/No), while the independent variables include age, gender, CGPA, sleep duration, profession, work pressure, academic pressure, study satisfaction, job satisfaction and dietary habits.

The study employs Support Vector Machines (SVM), Neural Networks, Logistic Regression, k-Nearest Neighbors (kNN), and Gradient Boosting to predict depression. The results indicate that Logistic Regression and SVM models provide the best generalization capability. In contrast, Neural Networks and Gradient Boosting models exhibit high accuracy on training data but suffer from overfitting, leading to decreased performance on test data. This study provides a novel approach to predicting depression among university students using machine learning techniques and contributes to data-driven strategies aimed at improving student mental health.