Implementasi Algoritma Ratcliffe/Obershelp pada Sistem Penilaian Otomatis Soal Uraian

Abstract

The use of electronic learning systems (e-learning) is being implemented at every level of education. From primary education to higher education. One of the applications carried out is the implementation of an examination system. The exam consists of several types such as multiple choice, description, similarity search, short answer, true/false, and calculation. The description exam itself consists of two types, namely the subjective description exam and the objective description exam. A subjective description test focuses more on a person's opinion regarding a condition or event, while an objective description test is a description test that has a definite answer. Almost all e-learning systems in Indonesia do not have the ability to automatically check answers to objective essay exams. Therefore, a system is needed to carry out automatic assessments on objective description questions. The automatic assessment in this research uses the application of the Sastrawi algorithm to find the origin of words and the Winnowing algorithm to perform synonym recognition to help the Ratcliffe/Obershelp algorithm determine the accuracy value obtained. This research data comes from an exam taken by 25 students, which consists of 4 questions and a test taken by 12 students, which consists of 5 questions. This system has ann average difference of 7,7 points compared to the assessment carried out manually.