A Framework for Source Code Plagiarism Detection

Abstract

Plagiarism has been a concern in the academia over the years and plagiarism in source code is now becoming a major concern with the recent advent of online source code portals and repositories. The paper presents an attempt to explore the ways of finding similarities between program source codes and also presents a framework for developers to check source code plagiarism. We discuss common ways of imitating others‟ source codes and the existing methods of plagiarism finding in source codes. Several approaches to find plagiarism have ended up on a tool and we briefly explain the capacity of mostly available and regularly used tools. We present a feature-wise comparison, a review on used technologies and a discussion of pros and cons of them. We introduce an algorithm called “Method Struct Comparer”. It is a method structure comparison algorithm based on attribute-counting and structure based technique. This also introduces how to implement a source code plagiarism checker using the plagiarism checking framework. The framework involves a traditional information Retrieving method called Vector Space Model and a string comparison algorithm called Karp-Rabin string matching algorithm. Our ultimate goal is to implement a tool for find plagiarism checking in C#.NET source code. The implemented tool has been tested with few corpuses. One corpus consists of more than one hundred undergraduate assignments of the UCSC. The corpus was able to unveil certain interesting results regarding the source code imitation. Keywords: Source code plagiarism, plagiarism checking framework, C#.NET, Plagiarism checking tool.