南昌工程学院学报2006，Vol.25Issue(2)：112-116,121,6.
Research on Application of Enhanced Neural Networks in Software Risk Analysis
Research on Application of Enhanced Neural Networks in Software Risk Analysis
摘要
Abstract
This paper puts forward a risk analysis model for software projects using enranced neural networks. The data for analysis are acquired through questionnaires from real software projects. To solve the multicollinearity in software risks, the method of principal components analysis is adopted in the model to enhance network stability. To solve uncertainty of the neural networks structure and the uncertainty of the initial weights, genetic algorithms is employed. The experimental result reveals that the precision of software risk analysis can be improved by using the erhanced neural networks model.关键词
software risk analysis/principal components analysis/back propagation neural networks/genetic algorithmsKey words
software risk analysis/principal components analysis/back propagation neural networks/genetic algorithms分类
信息技术与安全科学引用本文复制引用
Zhenbang Rong,Juhua Chen,Mei Liu,Yong Hu..Research on Application of Enhanced Neural Networks in Software Risk Analysis[J].南昌工程学院学报,2006,25(2):112-116,121,6.基金项目
Guang Dong Nature Science Foundation (04009863) （04009863）
