A Hybrid Framework for Software Clone Detection

Abstract

Software systems are becoming increasingly complex, and managing them is a critical topic
in the software business today. Cloning is one of the fundamental factors that makes software
maintenance more difficult. Code cloning is a copy-paste approach for reproducing code
portions. Developers find it simple to copy and paste code throughout the early stages of a
project's development. This paper presents a hybrid framework for detecting code clones. To
detect code clones, it combines a token-based technique with metrics-based technique.
Initially, metric based technique is applied to locate prospective clones. After identifying
prospective clones, token-based comparison is done to confirm that they are indeed clones.