Uncertain measurement for student performance evaluation based on selection of boosted fuzzy rules

Evaluation of student performance is one of the essential parts of educational systems that incorporated with uncertainty in human knowledge. Rating and predicting students’ educational achievement using arithmetical and statistical procedures may not explicitly advise the best route to estimate human acquisition of knowledge and experiences. Researchers in this field tended to adopt soft computing to defeat the challenges of analytical techniques in order to resolve this evaluation performance problem. Fuzzy logic is utilised to manage the intrinsic uncertainty associated with teachers’ subjective assessments and permits reproduction of student modelling in the linguistic form – the same form the human teachers do. The obstacle with existing fuzzy rule‐based systems is that the size of the rule base (the number of rules) grows exponentially with the expansion of the number of fuzzy sets included in the rules. This exponential increment in the size of the rule base enlarges the search time and the memory space needed. In this study, a fuzzy rule base compaction (densification) using genetic algorithm is suggested for uncertain measurement of student's performance. The proposed approach consists of three stages: knowledge recovery, coding, and compaction or optimisation. The comparative outcomes demonstrated that the recommended evaluator has a magnificent dynamic response than that of the traditional one.