Computerized adaptive test based on sugeno fuzzy inference system
Wrastawa Ridwan, Ifan Wiranto, Rahmat D.R. Dako

Electrical Engineering Department, State University of Gorontalo

Abstract

Along with the development of information and communication technology, assessment of student learning outcomes is no longer carried out in the form of written examinations, but rather carried out with a Computerized Adaptive Test (CAT). CAT consists of five main components, namely item bank, starting point, item selection, grading, and item completion point. CAT is adaptive because it allows the items given are selected according to the ability of students. Therefore, the CAT needs a method to estimate student ability. This research aims to design a Sugeno Fuzzy Inference System (SFIS) to estimate student ability. This fuzzy system consists of twelve IF-THEN rules, with four inputs, namely the students answer, the probability of students being able to answer correctly, the level of difficulty and discrimination of the questions. The output is an estimated ability, divided into five levels, namely very low, low, average, great, and excellent. Fuzzy system simulation is performed on linear algebra course with multiple-choice questions. The test will stop if the estimated value of the students ability does not change from the previous item. The simulation results show that the SFIS can estimate the ability of students by working on a maximum of six items each subject.

Keywords: computerized adaptive test (CAT), ability estimate, sugeno fuzzy system

Topic: Computer Science