Mutation Testing-Based Evaluation Framework for Evaluating Software Clone Detection Tools

Abstract

Mutation testing has become a prominent research area in the past few decades. The mutation testing has been basically used in the testing society. It is a type of software testing where we mutate (small change, modification in the program) source code using mutant operators by introducing potential new bugs in the program code without changing its behavior. Analogously, mutant operators generate new clones by copy/paste editing activities. However, several software clone detection tools and techniques have been introduced by numerous scientists and a large number of tools comprises for a perceivable evaluation. Moreover, there have been a lot of efforts to empirically assess and analyze variant state-of-the-art tools. The current abstraction exhibits that various aspects that could leverage the legitimacy of the outcome of such assessment have been roughly anticipated due to lack of legitimized software clone benchmark. In this paper, we present a mutation testing-based automatic evaluation structure for valuating software clone detection tools and techniques. The proposed framework uses the edit-based taxonomy of mutation operator for assessing code clone detection tools. The proposed structure injects software clones in the source code automatically, and after that, we evaluate clone detection tools. The clone detection tools are evaluated on the basis of precision (number of corrected clones) and recall (total number of clones). We visualize that such a framework will present a valuable augmentation to the research community.