Measuring Similarity Between Data Structures for Detecting Plagiarized Source Codes

Abstract

A program consists of data structures and algorithms. However, most studies, up to now, for detecting plagiarism of source codes are suggesting lopsided analyses considering only the algorithms (or instructions) of the source codes. This paper introduces a method for measuring the similarity between data structures for detecting plagiarized source codes. The proposed method was experimented with test sets including plagiarized source codes. The experimental result shows that the similarities among the data structures of plagiarized source codes are high degree as expected. This result implies that the similarity on data structures, along with the similarity on algorithms, is also one of the main factors to the decrease false alarms by lowering the threshold for the plagiarism.