Proposing a Model for Plagiarism Detection in Source Codes of Programming Language

Proposing a Model for Plagiarism Detection in Source Codes of Programming Language

Abstract—Source code re-use detection has become of growing significance as a common plagiarism prevention practice in the academic research. This paper describes a structure-based approach developed to detect Source Code Re-use (SOCO) in reference programs. The proposed model consists of three phases: preprocessing, implementing similarity metrics, ranking and decision-making. First, all codes are converted to a sequence of tokens as a specific preprocessing task. Then, a sequence alignment of instructions are invoked and constructed on behalf of the source code. Then, another approach is applied to the resulting sequence; and in the finishing phase, we subsequently report the result by the use of similarity measures to calculate the similarity values between the documents. To evaluate the method, we use the SOCO-2014 corpus. The comparison made between the submitted runs and the results obtained by the participants in the competition looks promising and thus, it seems to be the efficient and acceptable performance of the model proposed.