Plagiarism detection based on citing sentences

Abstract

Plagiarism, which is one of the forms of academic misconducts, is problematic. It results in discouraging innovation, and losing trust in the academic community. We modeled the plagiarism for academic publications, by means of the similarity between textual contents, and citation relations. Furthermore, we adopted the model in our proposed method for plagiarism detection. We evaluate our method using two types of dataset, namely auto-simulated and manually judged dataset. Our experiment shows that our method outperforms the baseline, which only uses the similarity between textual contents, on the auto-simulated dataset and the manually judged one for the ACL sub-dataset.