You are here:
(30 May 2011,
Date: 2011-05-27 Time: 11:00 Room: [[http://www.cs.uu.nl/docs/reach/bbl.php][BBL]] [[http://www.cs.uu.nl/info/plan/bbl.php]] ----+++++ Speaker: Marnix Kammer ----+++++ Title: Graph Algorithmic Techniques for Plagiarism Detection ----+++++ Abstract Ongoing research in the area of plagiarism detection at Utrecht University has indicated the need for an automated tool that can help identify possible cases of plagiarism in Haskell programs. There are many modiﬁcations that can be applied to the source code of a program in order to hide or mask an attempt at plagiarism. Existing tools like Holmes and MOSS make use of token stream analysis and document fingerprinting techniques to try to see through this, but call graph matching techniques are a relatively unexplored topic in this area. We have developed a tool based on a tree search algorithm that calculates graph edit distance in order to measure call graph similarity. Although calculating graph edit distance is an NP-complete problem and the resulting search tree constructed by our algorithm is exponential in size, we use preprocessing and domain-specific optimizations to be able to use the algorithm in practice. With experiments on fabricated cases of plagiarism and on actual student submissions for a Hask ell programming assignment we demonstrate the eﬀectiveness and usefulness of our tool in ﬁnding possible cases of plagiarism, in addition to the Holmes tool. In particular, when compared to the techniques used by the Holmes tool it is shown that our tool is less sensitive to many refactoring techniques that are typically used by plagiarists.
ore topic actions
Topic revision: r1 - 30 May 2011,
Copyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding UUCS?