DOCUMENT PLAGIARISM DETECTION USING DAMERAU LEVENSHTEIN ALGORITHM AND QUERY EXPANSION

Abstract

Plagiarism is considered a criminal act provided in the Law of the Republic of Indonesia Number 19 Year 2002 about Copyright, therefore plagiarism activities should be avoided. Computerized plagiarism detection needs to be done to reduce plagiarism against other people's work rapidly. One of the plagiarist's efforts in avoiding the existing plagiarism detection application is manipulating documents by replacing many words with synonyms. This research propose a design of plagiarism detection application by considering synonyms of words in documents, so that it can recognize documents even though the content is different, but the context is same. From the experiments that has been conducted, by comparing Damerau Levenshtein and Levenshtein Algorithm, the similarity values are relative similar for the handling of test case overall. But the handling of the case documents for typographical errors, Damerau Levenshtein Distance can recognize plagiarism documents better indicated by a higher similarity value 77,32%. And after integrated with query expansion, it can get significant result by detecting synonym words, so the application can detect document plagiarism more quite well Plagiarism is considered a criminal act provided in the Law of the Republic of Indonesia Number 19 Year 2002 about Copyright, t herefore plagiarism activities should be avoided. Computerized plagiarism detection needs to be done to reduce plagiarism against other people's work rapidly . One of the plagiarist's efforts in avoiding the existing plagiarism detection application is manipulating documents by replacing many words with synonyms. This research propose a design of plagiarism detection application by considering synonyms of words in documents, so that it can recognize documents even though the content is different, but the context is same. From the experiments that has been conducted, by comparing Damerau Levenshtein and Levenshtein Algorithm, the similarity values are relative similar for the handling of test case overall. But the handling of the case documents for typographical errors, Damerau Levenshtein Distance can recognize plagiarism documents better indicated by a higher similarity value 77,32%. And after integrated with query expansion, it can get significant result by detecting synonym words, so the application can detect document plagiarism more quite well