Enhancing Student Academic Performance Forecasting in Technical Education: A Cutting-edge Hybrid Fusion Method

Forecasting early-stage student performance within higher education is important to the academic community, offering a proactive framework to mitigate student attrition. However, gauging and prognosticating students' achievements in the Indian context are beset by formidable challenges due to the vast student populace and the deeply entrenched educational system. Each institution in India employs distinct criteria to assess student progress, lacking a standardized mechanism to oversee and appraise developmental trajectories. The past decade has witnessed diverse exploration of machine learning methodologies in educational research. Nonetheless, student performance prediction grapples with substantial obstacles, particularly when contending with imbalanced datasets. This research work adopts a dual-phase methodology to grapple with this quandary. Initially, conventional classification algorithms are deployed on a dataset encompassing the academic journeys of 4424 students. Subsequently, innovative hybrid machine learning (ML) algorithms are harnessed to yield more refined prognostications. The outcomes furnished by the proposed model furnish a platform for informed early decision-making of the advancement of higher education institutions. This streamlines the prediction of students' performance and empowers the educational domain to tackle these challenges with a more robust and insightful approach.