跳到主要导航 跳到搜索 跳到主要内容

Software defect prediction based on class-Association rules

  • Beihang University

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

Although there have lots of studies on using static code attributes to identify defective software modules, there still have many challenges. For instance, it is difficult to implement the Apriori-Type algorithm to predict defects by learning from an imbalanced dataset. For more accurate and understandable defect prediction, a novel approach based on class-Association rules algorithm is proposed. Class-Association rules are looked as a separate class label, which is a specific type of association rules that explores the relationship between attributes and categories. In an empirical comparison with four datasets, the novel approach is superior to other four classification techniques and accordingly, proved it's valuable for defect prediction.

源语言英语
主期刊名2017 2nd International Conference on Reliability Systems Engineering, ICRSE 2017
编辑Dongming Fan, Jun Yang, Ziyao Wang, Tingdi Zhao
出版商Institute of Electrical and Electronics Engineers Inc.
ISBN(电子版)9781538609187
DOI
出版状态已出版 - 8 9月 2017
活动2nd International Conference on Reliability Systems Engineering, ICRSE 2017 - Huairou, Beijing, 中国
期限: 10 7月 201712 7月 2017

出版系列

姓名2017 2nd International Conference on Reliability Systems Engineering, ICRSE 2017

会议

会议2nd International Conference on Reliability Systems Engineering, ICRSE 2017
国家/地区中国
Huairou, Beijing
时期10/07/1712/07/17

学术指纹

探究 'Software defect prediction based on class-Association rules' 的科研主题。它们共同构成独一无二的学术指纹。

引用此