Measuring of Scientific Document Abstraction Similarity Using Rabin-Karp and Poter Stemmer

Plagiarism is a plausible occurrence within the academic sphere. An evident instance involves the unauthorized replication of assignments in educational settings. To counteract such instances, there has been a contemporary advancement in systems aimed at curbing plagiarism. Nonetheless, subscribing to these systems for a fixed duration comes at a substantial cost. Hence, there arises a necessity for a freely accessible plagiarism detection system that is open to all individuals. In developing the system, of course, modeling must be done. in this study, it is proposed to use stemming Porter stemmer to overcome bi-language problems. Then for the similarity score model used Rabin-Karp and Porter Stemmer. The results will be compared with models in previous studies. Based on research that has been done using 50 sample data. The Rabin-Karp + Porter Stemmer model can perform better performance than the Rabin-Karp + Sastrawi Stemmer model, with a difference in similarity values of around 2-3 per cent. This result is also a significant difference based on the results of statistical testing using a t-test with an alpha value of 0.05. Further experimental research can be carried out using different models, such as winnowing and other models.