Multi-Agent based Sequence Algorithm for Detecting Plagiarism and Clones in Java Source Code using Abstract Syntax Tree

Abstract

Plagiarism and clone detection plays an important role in software security protection, software maintenance and license issues. Source-code similarity detection method can be classified as string-based, token-based, parse-tree-based and program-dependency-based. All of these approaches have certain limitations and can not meet the requirements when the source code is large and may produce false positives. But, parse-tree based detection improves the detection ability and efficiency. This paper describes method and statement based source code similarity detection, which detects the simple plagiarized code like exact match, near exact match and longest common sequence using multi-agent based detection which will perform the detection automatically. Automatic plagiarism detection will be helpful for code clone detection in software industry and plagiarism detection in projects.