String Matching based Plagiarism Detection for Document in Bahasa Indonesia

Abstract

One method to prevent plagiarism is to rely on plagiarism checker application. However, prevention of plagiarism to date in Indonesia still relies on the plagisarism checker held by institutions/organizations outside Indonesia for example Turnitin, Dupli Checker, Copyleaks, PaperRater, gramarly, etc. Plagiarism checker applications currently use methods that are applied for English. Development of plagiarism checker application for Bahasa requires support from multiple disciplines. Bahasa has a very unique pattern and different from English. String matching is an approach for plagiarism detection in computer sience. String matching approach is a “character by character” matching method. This approach could also use hashing block of the character and then use n-Gram to match the hashing block. But before n-Gram, file or ducument need a preprocesing such as tokenization, stopword removal, stemming, dan hashing with Rabin-Karp algorithm. The plagiarism detection system with a string matching approach generates a percentage of the similarity of the document by calculating the N-Gram result with Dice's Similarity Coefficient. For text extraction on a system built with tokenization, stopword removal, and stemming. The text then randomized using the Rabin-Karp method.