Research Information System
ACTA INFOLOGICA, vol.6, no.1, pp.127-139, 2022 (ESCI)
The rapid developments in technology have also been reflected in education. Educational institutions have started to support their computer-aided trainings with online and offline platforms. Exam evaluations can be carried out quickly by the system using these platforms. Meanwhile, exams containing test, classical, matching, and true-false types of questions are also applied to students in the classroom environment using traditional methods. The evaluation process for these takes a long time, especially when an exam has more than just multiple-choice questions. In this context, carrying out evaluations to be made accurately, reliably, and quickly becomes possible with the intelligent expert systems being designed. Using the word similarity weights algorithm (WSWA) and image processing techniques, this study develops software for quickly and reliably evaluating the traditional exams held in classrooms. By means of the proposed software, question types are separated, answers are distinguished from questions, handwriting is identified, and answers are evaluated on the exam papers using image processing techniques and WSWA. As a result of the experimental studies, the evaluation of just one exam paper is carried out in seconds, while the collective evaluation of all the papers in the exam is carried out in approximately 4 seconds per paper being evaluated. Thanks to this software, traditional exams can be evaluated quickly, efficiently, and accurately.