Code plagiarism detection based on graph density clustering

Code plagiarism has seriously endangered the healthy and orderly development of the software industry. Therefore, scholars and experts at home and abroad have proposed various types of code plagiarism detection technologies for this problem. In this paper, a code plagiarism detection method based on the graph density clustering algorithm is proposed to solve the problem of plagiarism in students’ programming assignments. In the proposed algorithm, the program dependency graph is applied to achieve the representative source code; Moreover, one-hot encoding is utilized to generate feature vector from the program dependency graph; Finally, Density-Based Spatial Clustering of Applications with Noise works as the clustering algorithm to achieve the code plagiarism detection. To verify the feasibility and effectiveness of the proposed approach, experimental is designed based on real programming assignments code datasets. Compared with some detection methods, experimental results show that the proposed algorithm based on graph density clustering has improved almost 10% in accuracy and has better time efficiency.