THE COMPARISON OF ALGORITHMS USED FOR PLAGIARISM DETECTION IN PROGRAMMING ASSESSMENTS

This paper examines available algorithms for plagiarism detection in the program source codes, used in the anti-plagiarism software. The instructors of the programming study course at the Department of Software Engineering struggle with increased plagiarism attempts in students’ homework; therefore, reliable plagiarism identification approach is required. Various researches suggest different solutions on the plagiarism detection topic, as well as main algorithms, used in their plagiarism detection software implementations. The main focus of this paper is to identify the algorithm ability to withstand source code automated alteration by an incompetent person applying modern Integrated Development Environment aids such as refactoring. The results of the present paper suggest that none of the conducted studies, including suggested algorithms, can fully cover all possible automated code changes to detect on-purpose induced changes which aim was to cheat anti-plagiarism detection. Further research will be focused on a different approach to solve the problem – higher abstraction level data usage for identification – abstract syntax tree and code meta-model generation, comparison and metric weight system development for assessment similarity estimation.