Enhancing BSIT Track Selection Using Predictive Modeling: A Multi-Class Classification Approach

Abstract

In today's evolving educational landscape, aligning students' skills with industry demands is essential. This study employs predictive modeling to improve the Bachelor of Science in Information Technology (BSIT) track selection at Bulacan State University. The BSIT program includes foundational courses in the first two years, followed by specialized tracks — Business Analytics (BA), Service Management (SM), and Web and Mobile Application Development (WMAD) — in the final two years. The study aimed to determine the characteristics of students who completed the requisite course requirements per track and graduated within the stipulated timeframe. It analyzed 481 student records admitted to the BSIT program during the 2019-2020 academic year, divided into a training set and a test set. Using the Cross-Industry Standard Process for Data Mining (CRISP-DM) and data-driven methods, this research recommends suitable specialization tracks - BA, SM, or WMAD - based on students' demographic and academic profiles. By analyzing historical data from timely graduates, significant attributes influencing track selection were identified. The study addresses class imbalance using the Synthetic Minority Over-sampling Technique (SMOTE) and evaluates model performance through accuracy, kappa statistics, and confusion matrix. The developed model is incorporated into a prototype prediction system, enhancing decision-making for personalized educational pathways and supporting timely graduations.