期刊文献+

基于主题建模的软件类代码可理解性评估方法

Evaluation Method of Software Code Comprehensibility Based on Topic Modeling
下载PDF
导出
摘要 代码可理解性是软件质量评价中必不可缺的要素之一,然而传统研究并没有对代码的可理解性进行量化评估。提出了一种基于LDA主题建模的代码可理解性评估方法,结合使用WordNet相似度计算算法和K均值聚类算法,实现了软件类代码可理解性的量化评估。选用Tomcat软件内的Catalina下core包内的40个Java类文件进行了实验,实验结果表明所提出的方法可以清晰地区分类代码的可理解性程度,并得到可靠的软件类代码可理解性评估结果。 Code intelligibility is one of the indispensable elements in software quality evaluation.However,traditional research has not quantitatively evaluated code intelligibility.A code intelli-gibility evaluation method based on LDA topic modeling is proposed.Combined with WordNet similarity calculation algorithm and K-means clustering algorithm,the quantitative evaluation of software code intelligibility is realized.40 Java class files in Catalina core package in Tomcat software are selected for experiments.The experiments results show that the proposed method can clearly distinguish the intelligibility of classification code,and obtain reliable evaluation results of software class code intelligibility.
作者 刘务 杨春晖 林军 袭安 LIU Wu;YANG Chunhui;LIN Jun;XI An(CEPREI,Guangzhou 511370,China;Ministry of Industry and Information Technology Key Laboratory of Industrial Software Engineering Application Technology,Guangzhou 511370,China)
出处 《电子产品可靠性与环境试验》 2023年第1期13-19,共7页 Electronic Product Reliability and Environmental Testing
关键词 代码可理解性 LDA主题建模 K均值聚类算法 评估方法 code intelligibility LDA topic modeling K-means clustering algorithm evaluation method
  • 相关文献

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部