Developing a Plagiarism Detecting Method of Paraphrasing Sentences by Syntactical Structure and Semantic Similarity
2012
Hochschulschrift
Zugriff:
100
Along with globalization of higher education in Taiwan, the contribution and development of academic researches become one of national economic development indicators. In order to enhance the competitiveness of universities in Taiwan, Ministry of Education pay more attention to the degree of internationalization in recent years. As a result, English has become the main medium for universities to publish academic researches. However, it is not an easy task for Taiwanese researchers─English as a foreign language learners─to compose a variety of writings in English. There is not enough English writing training for researchers to compose researches in English by themselves. With the advance of Internet, it is easier to obtain information using powerful search engines on the web. It leads researchers to copy the ideas of others intentionally or unintentionally. In fact, many researchers’ recognition of plagiarism is incorrect. They think that it is not regarded as plagiarism when they modify the documents slightly, so they does not recognize their behavior is illegal. Nevertheless, the major function of the currently available plagiarism-detecting softwares which plays the role of checking possibly plagiarized papers with the essay database or Internet-based search engines. These existing systems can only detect simple kinds of plagiarism based on only single terms. It causes that the cases of paraphrase can not be detected, and systems are not able to guide researchers to paraphrase properly. Hence, there is a requirement of an environment which can help researchers and educate them detect whether their documents are plagiarized or not. In this study, a new plagiarism detection method which can guide paraphasing the original sentences to avoid plagiarism will be proposed. The proposed method make use of the syntactical structure to retrieve all phrases of the sentences to improve the accuracy of using single terms. In addition, we considers the substitution and reversal of the terms, computing the semantic similarity and order similarity of the sentences. Find out the paraphrased documents which is suspected to avoid plagiarism through original document and user modified document. Finally, we are looking forward to suggest users how to avoid plagiarism. After evaluation, the proposed method can improve the accuracy of using single term traditionally. The semantic similarity can also take an advantage of the results. Moreover, the performance of using PATH and WUP to calculate the semantic similarity is better than only using PATH or WUP.
Titel: |
Developing a Plagiarism Detecting Method of Paraphrasing Sentences by Syntactical Structure and Semantic Similarity
|
---|---|
Autor/in / Beteiligte Person: | Ka-WengSou ; 蘇嘉穎 |
Link: | |
Veröffentlichung: | 2012 |
Medientyp: | Hochschulschrift |
Sonstiges: |
|