Implementation of Supervised Learning Algorithms for Early Detection of Depression in University Students
DOI:
https://doi.org/10.35730/ykrqbv55Keywords:
Depression Detection, Supervised Learning, University Students, Decision Tree, Support Vector MachineAbstract
Depression among university students is an escalating global mental health challenge driven by academic pressure, financial burdens, and psychosocial transitions. Early detection is critical to facilitating timely psychological intervention and preventing severe clinical outcomes. This study evaluates the performance of four baseline supervised machine learning classification algorithms—Logistic Regression (LR), Decision Tree (DT), Support Vector Machine (SVM), and K-Nearest Neighbors (KNN)—for the early detection of depression among university students. We utilized the Student Depression Detection Dataset comprising 4,464 student records and 12 mixed-type socio-demographic, academic, and lifestyle independent features alongside a binary target variable. Data preprocessing was executed via an automated pipeline incorporating StandardScaler for numerical features and OneHotEncoder for categorical attributes to prevent data leakage, followed by an 80:20 stratified train-test split (3,571 training and 893 testing records). Model performance was assessed using accuracy, recall (sensitivity), precision, and F1-score, with a primary clinical emphasis on maximizing recall to minimize false-negative screenings. Experimental results demonstrate that Support Vector Machine achieved the highest overall accuracy (98.88%), precision (99.15%), and F1-score (97.89%), whereas the Decision Tree classifier achieved the highest recall rate at 97.50% (detecting 234 out of 240 depressed students in the test set, with only 6 false negatives) and an accuracy of 98.54%. Logistic Regression and K-Nearest Neighbors yielded recall rates of 92.92% and 86.67%, respectively. These findings indicate that Decision Tree and Support Vector Machine models provide highly reliable, non-invasive screening frameworks capable of assisting higher education institutions in identifying at-risk students early
Downloads
Published
Issue
Section
License
Copyright (c) 2026 ilsahidayat, Irfan Kurniawan, Erwin Herian B (Author)

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.