Document Similarity Detection using Rabin-Karp and Cosine Similarity Algorithms

Plagiarism action often occurs in the world of education. One example of plagiarism in education, especially in the university environment, is the frequent occurrence of similarities between theses made by students. One solution is to develop an application that can detect similarities in thesis documents that can be used by all groups, especially to facilitate students. Natural Language Processing (NLP) is one of the fields that can be a solution in developing document similarity detection applications. In this study, testing focused on Indonesian documents. In addition, this study aims to build a system or application to detect the similarity of Indonesian documents using the Rabin-Karp algorithm with the Cosine Similarity method as a Distance-Based Similarity Measure and using the Stemming technique. The experiment results using 15 documents of scientific publications in Indonesian show that the Rabin-Karp algorithm with Cosine Similarity can be used to detect the similarity of published manuscripts, especially in Indonesian. With the fastest processing time of seconds and the longest around. seconds, this algorithm quickly detects document similarities, especially in documents. Further research is recommended to optimize from the algorithm side so that the processing time is faster with a much larger number of words.