Title :
Identification of nonlinear systems with missing data using stochastic neural network
Author :
Tanaka, Masahiro
Author_Institution :
Dept. of Inf. Technol., Okayama Univ., Japan
Abstract :
In this paper, nonlinear identification is dealt with by using Gaussian sum distribution. This model is also called a stochastic neural network. By using the stochastic model, it is possible to estimate the output and also the missing elements in the input vector within the framework of conditional estimation. The model parameters can be estimated by using the EM algorithm. By interpolating the unknown elements, we don´t have to discard the vectors including the missing elements
Keywords :
Gaussian distribution; neural nets; nonlinear dynamical systems; parameter estimation; EM algorithm; Gaussian sum distribution; conditional estimation; missing data; nonlinear identification; nonlinear systems; stochastic neural network; Equations; Interpolation; Kernel; Neural networks; Nonlinear systems; Parameter estimation; Stochastic processes; Stochastic systems; Training data; Yield estimation;
Conference_Titel :
Decision and Control, 1996., Proceedings of the 35th IEEE Conference on
Conference_Location :
Kobe
Print_ISBN :
0-7803-3590-2
DOI :
10.1109/CDC.1996.574577
