学术
摘 要:
阐述了神经网络集成的基本概念及其在液压故障诊断中应用时的个体网络生成方法和结论结合方法。根据液压系统的工作特点，采用Gauss随机函数作为个体网络的训练样本的随机采集函数，使个体神经网络的输出集中于各工作阶段的主要故障，并保持了个体网络之间的差异性，增强了网络集成的泛化能力。[著者文摘]
文章出处:
《现代防御技术》-2008年36卷1期,79 -45-49,79页
栏目信息:
文献标识码:
A
文章编号:
1009-086X（2008）-01-0045-05
Application of FNN Ensembles in Hydraulic Fault Diagnosis
NIE Guang-wei, QI Yi-hui Beijing Institute of Mechanical Equipment, Beijing 100854, China）
Abstract:
The basic concept of neural network vidual net and the combination of the conclusion in ensembles is introduced, the generation of the indineural network ensembles applied in fault diagnosis are also exposited. According to the working characteristic of hydraulic system, the gauss random func- tion is used to be as the random sampling function of the individual network train set. This method makes the neural network focus on the main fault and keeps the diversity of individual networks, the generalization ability is enhanced.[著者文摘]
Key words:
fault diagnosis; Bagging algorithm; neural network ensembles; fuzzy neural network （FNN） ; machine learning
更多>>免费文章
cqvip.com