JoWUA

Abstract

Plagiarism is an activity that often occurs in the field of education. Those who commit plagiarism can receive severe punishment, such as being expelled from the university or having their academic degrees revoked. Efforts to avoid plagiarism can be made through implementing a detection system developed over the last few years. However, the cost of subscribing to such a system for a particular time is relatively high. Besides that, another problem is limited access, where not all universities have subscriptions for such applications. Thus, we need a plagiarism detection system that all parties can access without charge. In developing this system, modelling must, of course, be carried out. In this research, performance comparisons were made between the Sastrawi+Rabin-Karp model, the Sastrawi+Winnowing model, the Porter+Rabin-Karp model and the Porter+Winnowing model. This comparison aims to determine the best model and then implement it in a system for detecting similarities in student assignments, essays and scientific documents. After testing the Porter+Winnowing model, it produces the best performance compared to the model based on the evaluation parameters of similarity score and processing time/execution time. This shows that the Porter+Winnowing model is more sensitive to the similarity of texts and documents. The Porter+Winnowing model can be developed into a document similarity detection system to help reduce the risk of plagiarism.