Lin Weixuan
ABSTRACT
This study developed a student academic performance early warning management system for Guangzhou Huashang Vocational College. The researcher used descriptive and developmental research methods to identify the current practices, difficulties, and challenges faced by the College through interviews and surveys with students, teachers, administrative departments, and education experts. By analyzing and evaluating student academic performance data using data mining techniques and algorithms such as decision trees and the XG Boost ensemble learning algorithm, an early warning model was established. The Software development model using SCRUM was adopted in the developmental design phase. The system allows real-time monitoring and timely warnings of students’ academic performance, providing personalized academic support measures. The goal is to improve student graduation rates through proactive prevention, process control, timely warnings, and academic support. A questionnaire was provided to IT experts who assessed the system’s compliance with ISO/IEC 25010 criteria, including functional suitability, performance efficiency, compatibility, usability, reliability, security, maintainability, and portability. The developed system was found to be “very” compliant with these criteria.
Keywords: Academic performance, academic warning model, academic assistance, xGboost algorithm, decision support system