Visualization of Plagiarism Detected in Documents

Tremendous amount of data is available digitally now a days and therefore need arises for a system which can automatically track any copying of data without permission. Copying of others copyright material is an illegal act. Plagiarism as it is called, is an act of copying others literary works and posing as though it is ones own work. Here in this paper we have developed a system which can automatically detect plagiarism and also uses the data to visualize the plagiarized document. A detector cell is created which will collect all the detected plagiarized information. Detector cell is created based on three criteria namely copying, reordering and replacing. Exact copy of the sentences are detected by checking sentence replication of certain sentences where keywords occur frequently. Reordering is possible by structurally modifying the sentences by changing the order of words. Replacement is done by replacing certain words with their synonyms in the copied sentences. As soon as the plagiarized sentences are detected the detector cell collects the information and updates itself for future references. Later the entire document is modeled as a plain three dimensional terrain surface over which the detector cell information is applied and therefore the plagiarized sentences appear as elevated surfaces and all other sentences appear as plain flat surface. The exact copy of sentences will have more elevation when compared to structurally reordered sentences and replacement sentences which are modeled as having less elevation. This modeling helps users in analyzing the document with respect to the amount of information copied, spatial location of copied sentences and the behaviour of plagiarism. This also helps in knowing the sentences which are plagiarized the most from multiple sources. The size of the document is also mapped to the size of the terrain surface so that spatial location information is maintained. Gestalt perception laws of similarity, proximity and continu- ity are also applied over the plagiarized data to obtain a gestalt based visual model.