A development and application of similarity detection methods for plagiarism of online reports

Recently, the network environment in classroom has been much improved. However, the facilities may promote the imprudent plagiarism of the other´s answer in online report submission. We are developing the detection tool for report plagiarism based on similarity. We adopt two kinds of approach for similarity judgment. The one is text base by editorial distance. Another is binary base by file compression ratio. We prepare some preprocess methods like morphological analysis and removing redundant information. We aim to realize adequate judgment precision and cost performance. The tool incorporated as a function of WebBinder, which is online report task support system. It uses not only for exposure of the fraudulent activity by a teacher but also for the warnings at the time of students´ uploading files. As preliminary experiment, we applied the prototype tool for similarity calculation and detection plagiarism. The samples are the three kinds of data; explanation text of information engineering term, mathematical document in binary file and C language source code.