Integrating Outcome-Based Education with Machine Learning Based Clustering to Enhance the Academic Support System for Slow Learners in Engineering Programs

Abstract

Beginning with how students perform, shifting toward outcome-focused teaching brings clarity through measurable goals per course. Rather than broad assessments, this work looks closely at first-year mechanical engineering pupils by tracking their progress across key classes. From semester one, records of 146 individuals spread over five main courses provided detailed insight, every class built around five specific objectives. Instead of overall grades, averages on these individual outcomes shaped a finer picture of each learner's grasp. On that foundation, sorting methods drawn from data science - grouping patterns and logic-driven rules - helped distinguish different types of performers. One cluster stood out early: those consistently below peers, later labeled as needing more time or help. In total, results split the batch into three sections - one small segment struggled, most held steady ground, while another group showed stronger command. Numbers ended up being 23 who learned slower, 89 fitting a middle range, alongside 34 showing advanced understanding. Later in Semester 2, mentoring and tailored academic help became available for students recognized as slower to grasp material. Performance indicators tied to course objectives - within an outcome-based education model - paired with cluster analysis of student data, helped shape individualized assistance that improved results among those struggling academically.