Hybrid System For Plagiarism Detection

Proceedings of Recent Advances in Natural Language Processing, pages 527–532, Hissar, Bulgaria, 12-14 September 2011. Hybrid System for Plagiarism Detection Javier R. Bru University of Alicante javier.r.bru@gmail.com Patricio Martínez-Barco University of Alicante patricio@dlsi.ua.es Rafael Muñoz University of Alicante rafael@dlsi.ua.es Abstract The Internet boom in recent years has in- creased the interest in the field of plagiarism detection. A lot of documents are published on the Net everyday and anyone can access and plagiarize them. Of course, checking all cases of plagiarism manually is an unfeasible task. Therefore, it is necessary to create new sys- tems that are able to automatically detect cases of plagiarism produced. In this paper, we in- troduce a new hybrid system for plagiarism detection which combines the advantages of the two main plagiarism detection techniques. This system consists of two analysis phases: the first phase uses an intrinsic detection tech- nique which dismisses much of the text, and the second phase employs an external detec- tion technique to identify the plagiarized text sections. With this combination we achieve a detection system which obtains accurate re- sults and is also faster thanks to the pre- filtering of the text. 1 Introduction Plagiarism detection is a topic that has always received some interest. Authors have worried about other people stealing their intellectual property, in other words, having their work pla- giarized. With the recent increase in the impor- tance of the Internet plagiarism has become a real problem. Anyone anywhere in the world can access any document, plagiarize and publish it as their own. Each author cannot spend all his or her time watching that nobody copies his or her work, so it is very important to create systems that can automatically detect cases of plagiarism. The research in this field is mainly divided in- to two branches: external plagiarism detection and intrinsic plagiarism detection. Each one has its own advantages and disadvantages. In this paper we introduce a new plagiarism detection system that combines these two detection tech- niques, joining their main advantages and avoid- ing their disadvantages. This system has a first phase that uses an intrinsic detection technique to identify text sections that are most likely to be plagiarism. This phase helps us to filter the text and discard much of it, thus the next phase must analyze less text. The second phase is based on an external detection technique, which employs text comparisons to identify plagiarized sections. This technique, although slow, is very precise for plagiarism detection. Moreover, the problem of slowness is mainly solved thanks to the filtering of text done in the previous phase. The benefits of this combination of detection techniques are the merge of the speed of intrinsic detection and the precision of external detection. We also avoid their disadvantages. In intrinsic detection we improve precision with the second analysis phase. About external detection, which is a very slow technique, we increase speed thanks to the filtering of text in the first phase. The remainder of this paper is organized …

Content truncated to 500 words