Fuzzy logic controller development for prediction of student performance based on activity parameters

Abstract

Owing to the COVID-19 pandemic, most of the academic education has suddenly shifted from traditional teaching methods to advanced technological methods on the internet. Many teachers encountered difficulties in successfully evaluating and monitoring their students. We address these challenges and propose a fuzzy logic-based controller that can assists teachers during classes and support allocation of appropriate resources to students. The purpose of the controller is to provide early warning about students who have performed poorly in the initial part of the course assessment. The controller makes predictions based on 5 input parameters which, by applying statistical tools, have been proven to accurately reflect the students' achievements. The model was tested on a group of 50 students and the results indicate 82% prediction accuracy. There is a possibility for additional improvements related to the built-in parameters, both in terms of their selection and in terms of their number.