To enhance the code clone detection algorithm by using hybrid approach for detection of code clones

Abstract

Code clones are easy and quick way to add some existing logic from one section to another section. Code clones however increase risk of bug multiplication with each copy of duplicate code if there is a bug in source of clone. Clone is a persistent form of software reuses that effect on maintenance of large software. In previous research, the researchers emphasize on detecting type 1, type 2, type 3 and type 4 type of clones. The existing code clone detection techniques are available like text based, token based, abstract syntax tree, program dependency graph and metric based technique are used to detect clone in source code. In this research, the enhancement in code clone detection algorithm has been proposed which detects code clones by HYBRID approach that is combination of program dependency graph and Metric based clone detection techniques. In this work, firstly implementation of code clone detection will be done by hybrid approach on various datasets. Then, comparison of existing technique will be done with the hybrid technique in terms of achieving enhancement in performance, efficiency and accuracy in results. This method is considered to be the least complex and is to provide a most accurate and efficient way of Clone Detection. The results obtained have been compared with an existing tool for the open source of web applications.