Personalization of Test Sheet Based on Bloom’s Taxonomy in E-Learning System Using Genetic Algorithm

Abstract

In E-learning systems during tutoring, evaluating the learning status of each learner is essential, and tests are a usual method for such evaluation. However, the quality of these test items depends upon the degree of difficulty, discrimination, and estimated time. While constructing the test sheet, the selection of appropriate test items is also important. This paper aims to provide a method to generate the test sheets based on the different learning levels of Bloom’s taxonomy using genetic algorithm. The questions are initially categorized into six different learning levels based on the keywords given by Bloom’s taxonomy. These six different learning levels are assigned difficulty degrees from 0.1 (lowest) to 0.6 (highest). Then, the different number of questions is generated for all these six different levels of learning using genetic algorithm. The numbers of questions are generated in such a manner that the total difficulty degree of all the questions is equal to the target difficulty degree given by the instructor. Based on the test sheet generated, learner performances are being analyzed, and by doing so, even weaker learners are able to attend low-level questions and perform well in the examination. This would also help to analyze at which level of learning learners are facing problem so that effort can be made to improve their learning status.