Evaluation of the accuracy of winnowing, rabin karp and knuth morris pratt algorithms in plagiarism detection applications

Abstract

The unethical behaviour of acts of plagiarism has been a disgrace in the educational realm. Using the internet, people can easily find articles or documents that are relevant to their current work, and simply duplicate the sentences or paragraphs without paraphrasing or giving correct citations. Such action falls into the area of plagiarism. In order to minimize the problem, especially in the educational field, it is necessary to develop plagiarism detection applications. The algorithm employed in the application plays an important role in obtaining accurate plagiarism detection results. To the best of our knowledge, three algorithms are commonly used in plagiarism detection applications, namely Winnowing, Rabin Karp and Knuth Morris Pratt, which are all employed in our application. To specify the accuracy of each algorithm, the percentages of the plagiarism detection results are compared to the results from examination by a human expert. From our results, we found that the order of the accuracy from highest to lowest corresponded to the Winnowing algorithm, Rabin Karp algorithm and Knuth Morris Pratt algorithm, with value differences of 1.19%, 53.91% and 83.91% respectively.