Learning activities require supervision and attention from the teacher as a facilitator. In general, one of the learning media, e-learning, only able to carry out learning activities without more supervision from the teacher. This can encourage students to do things that should not be done in learning activities, such as plagiarizing or cheating when the exam is carried out. Junior High School of Situbondo 3 as a school that does not have e-learning until this research is carried out, has problems in supervising students during exams. It is not uncommon for students who have below-average achievements to get high score on exams. Based on this problem, the researchers developed an e-learning that is able to detect potential of cheating when online exam is carried out. The method used to detect this potential is the Semantic Textual Similarity (STS) method for essay type exams and Angoff's B-Index method for multiple choice exams. The implemented module produces a value of up to 0,92 on the STS method and the right detection on the Angoff B-Index method. The module also reduces the actions of potential cheating with the potential for occurrence of 11,54% and 19,52% of the two classes with the same conditions.
Copyrights © 2021