The Application of Computer Adaptive Test Using Expert System in E-Learning

Abstract

Higher education graduates must have competency abilities in accordance with the learning outcomes that have been formulated equivalent to the Indonesian National Qualifications Framework (KKNI) which is a problem that must be resolved at this time. According to the KKNI occupation map, a student must have skills in accordance with course learning outcomes. Where graduates in tertiary institutions are still difficult to compete in the world of work because they are not supported by the skills and competencies they got after graduating. One solution that can be used to help tertiary institutions in seeing student competencies that are easily accessible at affordable costs is through e-learning. The purpose of this research is to model the learning evaluation process by conducting online exams in the form of Computerized Adaptive Testing (CAT) using an expert system in the e-learning application, where test items are analyzed based on the level of difficulty, distinction, and function of the distractor. The procedure for selecting a test item for the diagnosis of thinking skills is based on the revised concept of Bloom's cognitive taxonomy and the expert system uses the rule-based method that provides a solution to the problem of measuring student competency abilities. The research can measure the competence of students by providing test questions that have a difficulty level. The CAT application can administer the items to be tested, grouped based on 3 levels of the test taker's thinking ability, namely questions with low, medium and high levels. Furthermore, the item calibration is entered into the CAT-assisted database using a rule-based expert system. Another facility of the CAT application is that it can display randomized questions by combining the Multistage Testing (MST) model in e-learning.