A benchmark study on the effectiveness of search-based data selection and feature selection for cross project defect prediction.

Information and Software Technology(2018)

引用 99|浏览74
暂无评分
摘要
•GIS significantly improves CPDP benchmarks based on multiple tests.•GIS has a comparable performance to WPDP benchmarks and it does not require even limited WP data.•Search based instance selection combined with feature selection can effectively improve defect prediction performances.
更多
查看译文
关键词
Cross project defect prediction,Search based optimization,Genetic algorithms,Instance selection,Training data selection
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