INDONESIAN TEXT DOCUMENT SIMILARITY DETECTION SYSTEM USING RABIN-KARP AND CONFIX-STRIPPING ALGORITHMS

摘要

Nowadays, negative impact, such as plagiarism, may arise along with fasterand easier ways in finding information. There are many software and websites that can beused to check the occurrence of plagiarism, but unfortunately, they are not really suitablefor scientific papers which are written in Bahasa Indonesia because it is designed for textin English. Therefore, a document similarity detection system that is more suitable forpapers written in Bahasa Indonesia is needed. Rabin-Karp is an algorithm that can beused in checking the similarity between documents, while Con fix-Stripping is an algorithmthat can perform basic word seo/rch in Bahasa Indonesia. This research has successfullyimplemented Rabin-Karp and Confix-Stripping algorithms very well. Tests performedwith various document scenarios as well as algorithms have given some performanceresults of the system in terms of time and similarity level. The system with the pureRabin-Karp can provide the best system performance, both in terms of time and accuracywith an average total processing time speed of 0.0123 second and the average similarityrate of 89.1967%. The accuracy level given by the system is 0.7. The system that hasbeen added with a stemming process or N-Gram can also improve some test results interms of processing time and similarity level.