Keyword: student participation

ABSTRACT: Enhancing students' academic performance in higher education is a primary goal, necessitating a systematic and adaptable approach. Traditional methods for analyzing student performance data often struggle with complexity and unpredictability, making them inadequate for handling intricate educational patterns. As a result, the development of fuzzy logic-based decision-making systems has become increasingly important. This study aims to design an accurate fuzzy logic system for predicting first-year physics students' academic achievement at Dire Dawa University. The model leverages the fuzzy library, defining membership functions for participation in experiments and discussions (categorized as low, medium, and high). It then establishes rules to map participation levels to predicted grades (poor, average, good, and excellent). The analysis involved applying the fuzzy logic system to a dataset and validating the predictions against actual grades. The findings revealed that the model accurately predicted grades for medium participation (e.g., 50.00 for 50% participation) but tended to overestimate high participation levels (e.g., predicting 85.00 when the actual score was 67.51). This research contributes to educational technology by providing a flexible predictive tool that can be expanded to other STEM disciplines, enhancing data-driven decision-making in academic settings. These results demonstrate the effectiveness of fuzzy logic in managing educational uncertainties, though refinements are needed to address overestimation and incorporate additional variables such as study habits and prior knowledge.