Word Sequence-based Newspaper Articles Plagiarism Detection

Currently, most media publishing companies publish their news in real time via the World Wide Web, which enables instant access to newspaper articles. In this rapid publishing environment, original new articles can be readily reproduced and partially updated. Plagiarism of newspaper articles violates timeliness, and can severely diminish the economic value of newspaper articles, particularly that of the original newspaper article. Until now, however, plagiarism between newspaper articles in the news industry has yet to be addressed thoroughly. In this paper, we propose a semi-automatic plagiarism detection method based on gene sequence alignment for Korean newspapers. In order to solve the problem of the difficulty inherent to manual examination, our plagiarism detection system automatically recommends plagiarism candidates to experts after processing, including stop-words elimination, morphological analysis, word sequence alignment, similarity computation, and plagiarism determination steps. Herein, we demonstrate experimentally that our system evidences reasonable precision and recall.