Analysis of Stationary Dynamic Neural Networks on The Basis of Statistical Linearization Method

摘要：

The approximate methods are proposed for determination of statistical moments of order 2 on the output of finite-memory dynamic neural networks and recurrent neural networks on the basis of similar characteristics of steady-state input process. The paper examines special features of application of statistical linearization method for analysis of the above-mentioned neural networks included in the closed-loop control contour.

作者: Vladimir A. Ptitchkin

作者单位: Department of Information Technologies and Control Belanissian State University of Informatics and Radioelectronics 6, P. Brovky Str. 220027 Minsk, Republic of Belarus

会议类型: 国际会议

会议名称: 8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)

会议地点: 上海

会议语种:英文

页码: 1582-1587

在线出版日期: 2001-11-14（万方平台首次上网日期，不代表论文的发表时间）