PLAGIARISM DETECTION SYSTEM FOR REPORTS BASED ON LARGE-SCALE DISTRIBUTED ENVIRONMENT USING IDLE COMPUTERS

摘要

We developed a plagiarism detection system for report examinations in university education. The system consists of two parts: a plagiarism detection function and a large-scale distributed environment. The distributed environment included 560 idle computers; hence, the construction cost was low. The system was applied to four report examinations at a university. Initially, about 40% of students produced reports containing plagiarized sentences. Following the application of the system, 10% of students stopped plagiarizing. Interviews with teachers showed that they appreciated the deterrent effect on plagiarism. An analysis of four sets of applications showed that general questions in report examinations led to an increase in plagiarized reports. Finally, we proposed a method for avoiding incorrect plagiarism detection in student reports. The accuracy of plagiarism detection was drastically improved. Thus, the results of the plagiarism detection system should prove useful for keeping the high quality of university education.