Ratcliff/Obershelp Algorithm as An Automatic Assessment on E-Learning

Abstract

There have been many major changes in the education sector due to the closure of educational institutions to suppress the spread of the Covid 19 pandemic, one of which is teaching methods in Indonesia. Problems arise when schools have to use e-learning for students and online assessments have to be done. It is known that difficulties are experienced when taking semester exams or when teachers give assignments online. This study applies text mining techniques using the Ratcliff/ Obershelp algorithm to determine the similarity value between two strings, namely between student answers and the teacher's answer key. A collection of answer data obtained from one of the 6th grade teachers at elementary school and applied to the Examz web application. This application is built to find out the status of answers based on the error tolerance that has been predetermined by the teacher. The results of this study indicate that the Ratcliff/ Obershelp algorithm has achieved a correction accuracy rate of 90% in Natural Science subjects, 93% in social science, 83% in Civil Education, 91% in SBK, and 97 in Religion. The final result for the average accuracy of the application is 91.00% in determining the status of student answers.