基于主成分分析和神经网络的软件项目风险评估
Risk Assessment of Software Project Based on Principal Component Analysis and Neural Network
随着软件行业的高速发展,软件项目的开发风险问题逐渐受到业界和学界的重视。由于软件项目风险评价存在典型的非线性特征,本文采用神经网络对软件项目进行风险评价;同时,由于软件项目风险影响因素众多、关联性大,本文利用主成分分析对数据进行降维处理,并将此方法实际应用于某软件公司的项目风险评价中,结果表明:主成分分析与神经网络结合使得神经网络收敛迅速、精度较高,评价模型性能良好。本研究拓展了软件项目风险管理的方法,丰富了神经网络学习与泛化能力。
With the development of software industry, the risk of software project attracts extensive attention increasingly within the industry. Due to there be feature, which is typical nonlinear of risk assessment of software project, therefore, the paper make risk evaluation by way of applying neural network. And meanwhile, there are so many related factors, which can affect risk of software project, here we simplify dimension of data by conduct principal component analysis. By study the case of a practical application in a software company, we found that, evaluation model predict accurately and has a high practical value by combining use of principal component analysis and neural network.
童磊、张晓航、侯超、崔然
计算技术、计算机技术
管理科学与工程软件项目风险主成分分析神经网络
risk of software projectthe principal component analysisneural network
童磊,张晓航,侯超,崔然.基于主成分分析和神经网络的软件项目风险评估[EB/OL].(2024-03-15)[2025-08-10].http://www.paper.edu.cn/releasepaper/content/202403-211.点此复制
评论