<![CDATA[Exams are One of the methods used in the teaching and learning process to evaluate the quality of learning. However, they also create additional workload for teachers, especially during grading, as it takes time to compare students’ varying answers, particularly for essay questions. The goal of this project is to create a web-based, real-time grading system that will help teachers in correcting essay answers quickly and easily. The system employs the Cosine Similarity method to measure similarity between student answers and the answer key based on text vector representations, as well as the Levenshtein Distance algorithm to calculate similarity between two strings through the fewest possible operations (substitution, deletion, and insertion). Mean Absolute Error (MAE) and Mean Absolute Percentage Error (MAPE) were used in testing to gauge the correctness of the system by comparing results with manual grading. The results showed an MAE value of 0.003298 and a MAPE of 0.002484%, indicating a very low error rate, meaning the system can produce scores close to actual values. Additionally, User Acceptance Testing (UAT) showed a high acceptance rate averaging 90.98%, indicating the system is suitable for use.]]>
Copyrights © 2026
