Using Kohonen Maps and Singular Value Decomposition for Plagiarism Detection

Plagiarism has become one area of interest for re-searchers due to its importance, and its fast growing rates. Effective clustering methods and faster search tools for matching and discovering the similarities between documents were the main two areas for the researchers. Many tools and techniques have been developed for plagiarism detection. In this paper we use singular value decomposition for its effective clustering of the documents in-order to reduce search time by creating a new matrix with fewer dimensions used for clustering the original (source) documents, and we use Neural Networks for local matching and comparison between a suspicious document and a source document, Kohonen maps (Self-organizing maps (SOM)) used to visualized and comparison of the result, in which represent the result as picture that easier to be analyzed.