基于主成分分析和BP神经网络的软件工作量估算方法

Effort Estimation Method Based on Principal Component Analysis and BP Neural Network

为构建适用不同公司的软件工作量估算模型,本文提出基于主成分分析和反向传播(Back Propagation,BP)神经网络的软件工作量估算方法。该方法采用主成分分析方法提取软件工作量的主要影响因素,并通过BP神经网络建立这些因素与软件工作量之间的映射关系。本文基于某公司的历史数据,采用该方法构建估算模型,并对其进行验证。实验结果证明,使用该方法构建的估算模型可有效地提高软件估算精度,实现软件项目合理的规划和科学的管理。

In order to construct the software effort estimation model suitable for different companies, this paper proposes a software effort estimation method based on principal component analysis and BP neural network. This method uses the principal component analysis method to extract the main influencing factors(feature attributes) of software workload, and establishes the mapping relationship between these factors and software workload through BP neural network. Based on the historical data of a company, this paper constructs the estimation model and verifies the method. It is proved that the estimation model constructed by this method can improve the accuracy of software estimation and realize the reasonable planning and scientific management of software project.