Detecting documents plagiarism using winnowing algorithm and k-gram method

Abstract

In this paper, we propose and evaluate a web-based software to check similarities of documents. The resemblance value of those documents will be compared based on the percentage of its word resemblance. The similarity value will help to detect plagiarism in documents. Methods used in this application are winnowing algorithm and web-based k-gram. We evaluate the accuracy of the system by comparing the system result with the human result. The differences between the systems and the respondents are 7% with k-gram 25 and 4% with k-gram 20. Moreover, processing time of our application are also discussed.