Toward plagiarism detection using citation networks

Abstract

Plagiarism makes people lose trust and discourages innovation in the research community. In this paper, we proposed a model for plagiarism using citation networks. More precisely, we used fragments of texts citing a document as surrogates for that document. Furthermore, we applied the model to a plagiarism detection system, and evaluated its effectiveness. Our experiment results suggest that our method may be able to detect plagiarism although the source documents of a plagiarized one are not included in a collection of documents. Moreover, our method improves the existing method when they are combined.