| 静态代码属性与软件模块缺陷数的偏相关分析 |
| |
| 引用本文： | 林开标,马樱,胡建强,朱顺痣.静态代码属性与软件模块缺陷数的偏相关分析[J].鹭江职业大学学报,2014(5):62-66. |
| |
| 作者姓名： | 林开标 马樱 胡建强 朱顺痣 |
| |
| 作者单位： | 厦门理工学院计算机与信息工程学院,福建 厦门361024 |
| |
| 基金项目： | 福建省教育厅科技项目（JA14234）;厦门理工学院高层次人才引进项目（YrO13027R）;厦门理工学院国家自然科学基金预研项目（XYK201432） |
| |
| 摘 要： | 为了解决因选择不同的静态代码属性子集而带来的不同软件缺陷预测性能的分歧，采用偏相关分析法，分析了静态代码属性之间的相关性，以及该相关性对预测算法的影响。在Eclipse数据集（包括发布前与发布后的缺陷数据）上的实验结果表明，静态代码属性之间存在偏相关。同时，也证实了偏相关是软件质量分析的非常重要的因素，为建立软件缺陷预测模型提供了可靠的分析方法。
|
| 关 键 词： | 缺陷预测 静态代码属性 实证 复杂度 |
Partial Correlation Analysis for Number of Defects in Software Modules and Static Code Attributes |
| |
| LIN Kai-biao,MA Ying,HU Jian-qiang,ZHU Shun-zhi.Partial Correlation Analysis for Number of Defects in Software Modules and Static Code Attributes[J].Journal of Lujiang University,2014(5):62-66. |
| |
| Authors: | LIN Kai-biao MA Ying HU Jian-qiang ZHU Shun-zhi |
| |
| Institution: | ( School of Computer ＆ nformation Engineering, Xiamen University of Technology, Xiamen 361024, China) |
| |
| Abstract: | This article analyzed the correlation between static code attributes and its effect upon predictor using partial correlation method in order to find a solution to the difference in the software defect predictors from using different static code attribute subsets. An experiment was performed using the public Eclipse dataset with both pre-release and post-release defects, and partial correlation was found to exist in the static code attributes. It has been proved to be a very important factor in software quality analysis, and provides a reliable analytical method to establish a model of the software defect predictors. |
| |
| Keywords: | defect prediction static code attributes empirical complexity |
| 本文献已被 维普 等数据库收录！ |
