摘要
目前,许多软件可靠性增长模型(SRGMs)被提出并应用于软件开发过程,但还没有在不同条件下都表现良好的普适性模型。将若干独立模型进行组合可提高单个模型的可靠性评估和预计精度。本文基于机器学习算法(Boosting算法),建立基于单个模型变异的动态赋权组合模型(ASCM)。ASCM模型可有效地改进单个原始模型的拟合性能。
At present, many software reliability growth models （SRGMs） have been proposed and applied to the software development process, but have not yet performed a good universal model under different conditions. Combining several independent models can improve the reliability and expected accuracy of a single model. Based on the machine learning algorithm （Boosting algorithm）, a dynamic weighting model （ASCM） based ona single model variation is established. The ASCM model can effectively improve the fitting performance of a single original model.
出处
《数字技术与应用》
2017年第8期120-123,共4页
Digital Technology & Application
关键词
软件可靠性
机器学习
动态赋权
组合模型
software reliability
machine learning
dynamic empowerment
combinatorial model
