Reconceptualizing examination debar criteria using fuzzy logic

Abstract

Strict requirements regarding student attendance have always been a debated topic in academic institutions. Numerous studies carried out to relate class attendance with a student’s overall performance have reported positive as well as negative results; thereby not resulting in a clear overall conclusion. Therefore, this paper presents a fuzzy logic based attendance evaluation system for higher educational institutions. The proposed fuzzy system considers four attributes: student attendance in the current course, overall performance, performance in the current course and faculty’s assessment for deciding if the student should be debarred from examination, allowed taking the examination or be given reconsideration. Since the considered attributes are relevant to any course, it results in a generalized model which may be adapted according to the specific requirements of courses at different universities. The proposed model is implemented using the fuzzy logic toolkit of OCTAVE. The application of the system to actual students’ data has yielded an accuracy of 95.25%. Further, for performance analysis, three classification algorithms, namely Naïve Bayes, Support Vector Machine and Neural Networks are also applied on the same dataset.