基于特征选择的软件缺陷预测方法

软件缺陷对软件功能的实现具有不可预知的危险,是软件产品的固有成分,提高软件的可靠性,关键在于降低软件缺陷出现的概率,而如何利用已有缺陷数据构建预测系统框架是研究的重点.针对传统软件测试技术虽然能够有效发现软件缺陷,但需要消耗大量的时间和精力,制约软件开发效率的缺点,提出基于特征选择的软件缺陷预测方法,算法对软件缺陷模型的经验数据集进行多特征选择,进而克服数据集之间的冗余性移除无关特征,得到缺陷模型的分类,最终实现软件缺陷的精确预测.实验表明,基于特征选择的软件缺陷预测方法具有较好的预测效果和较高的应用价值.

Research on Software Defect Prediction Method Based on Feature Selection

Software defects have unpredictable dangers to the implementation of software functions,and are inherent components of software products. The key to improving the reliability of software lies in reducing the probability of software defects,and how to use the existing defect data to construct the prediction system framework is the focus of research. Although the traditional software testing technology can effectively find software defects,it needs a lot of time and effort so as to limit software development efficiency. So this paper proposes a software defect prediction method based on multi-feature selection. The algorithm performs multi-feature selection on the empirical data set of the software defect model,and overcomes the redundant removal irrelevant features between the data sets,and obtains the classification of the defect model,and finally realizes the accurate prediction of software defects. Experiments show that the software defect prediction method based on multi-feature selection has better prediction effect and higher application value.