Code Similarity Detection by Computing Block Level Feature Metrics

All software systems are subject to continuous evolution and maintenance activities in order to eliminate defects and extend their functionalities. We need to deal with code duplication in order to prevent some problems that will appear when trying to adapt to the changes that are imminent in a real software system. In this paper, source code is divided into different types of blocks to detect duplicate segments. The proposed method aims at detecting segments that have similar functional blocks. We have experimentally validated the proposed technique in terms of accuracy. The method can efficiently identify structural and functional similar clones.