Needle: detecting code plagiarism on student submissions

Abstract

Code plagiarism is one of the most prevalent academic dishonesty activities in programming practicums. Automated code plagiarism detection plays an important role in preventing plagiarism and maintaining the academic integrity. This paper describes a novel code plagiarism detection algorithm needle, which is based on the network-flow approximation of editing distances between programs. This paper also presents the effectiveness and efficiency evaluation of the algorithm, the lessons and experiences learned from applying needle in practice, and discussions of the future challenges.