MTCD: An Efficient Cloning Detection Technique Based on Method Table.

TASE(2023)

引用 0|浏览2
暂无评分
摘要
Code cloning is an effective approach to accelerate software development, but may leave potentially significant pitfalls for maintenance. State-of-the-art clone detection techniques rely heavily on accurate syntax and semantic features which are difficult to identify and extract, making them hard to deploy in practice. This paper presents a simple but efficient approach called MTCD which extracts from a Java method (in its Java class file) the method table section of the code as its feature, and then transforms the feature into a decimal series for comparison. MTCD significantly simplifies feature identification and extraction. Experimental results on BigCloneBench suggested that MTCD clearly outperformed other established approaches by achieving an average F1-score of 0.95 with a modest training cost of 181 s for a 60 k dataset.
更多
查看译文
关键词
efficient cloning detection technique
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要