Exploiting Document Similarities for Plagiarism Detection

• Main Authors: Heng-rui Zhang, 張恒瑞
• Other Authors: Wei-Guang Teng
• Format: Others
• Language: en_US
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/19494832415803845710

碩士 === 國立成功大學 === 工程科學系碩博士班 === 95 === As information and networking technologies advance, people can easily get what they need on the web. This facilitates the learning and sharing processes among people. However, the plagiarism problem is also becoming more and more serious if people depreciate the creativity and intellectual property of others. An effective way to reduce the impacts of plagiarism lies on the detection techniques. In this work, we focus on extending the capabilities of identifying document similarities for plagiarism detection. Specifically, two crucial issues are addressed in this thesis. The first issue is on devising a proper technique to segment a suspicious document into smaller pieces for following steps to identify possibly multiple sources. On the other hand, since a plagiarist may slightly revise the grabbed contents when compiling into the plagiarized document, a technique to identify partial changes in a text segment should be developed. Moreover, our approach is carefully designed to reduce redundant computation cost when conducting comparison of document similarities. To verify the feasibility of our approach, empirical studies show that plagiarized documents and thus the malicious users can be precisely identified in a very efficient way.