Students’ Learning Performance Evaluation Using a New Fuzzy Inference System

Abstract

The research of computer technology in the field of education makes Intelligent Tutoring System (ITS) birth and develop rapidly. The main advantage of an ITS is it can provide appropriate teaching content according to students’ knowledge state, level and ability. Reasonable evaluation of students’ learning performance is the key and difficult point in ITS design. Using fuzzy logic to evaluate students’ learning performance is not a very new method. However, most approaches rely extremely on student’s answer time for each question, which are seldom collected in common exams. Given this insight, we present a new fuzzy inference mechanism in this paper to calculate students’ ranking order without dependence of the answer times. According to the importance, complexity and difficulty of the problem, new reasoning rules are set up to accurately evaluate the comprehensive level of problem. The results of the case study show that each student’s grades are distinguished, even for those with the same grades. The proposed method in this paper enriches the present evaluation approach of students’ performance, and is utilizable to analyze the direct relationship of students’ scores with the attributes of each question.