State space neural network. Properties and application

Zamarreño, Jesús M.; Vega, Pastora

doi:10.1016/s0893-6080(98)00074-4

Cited by 75 publications

(34 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…, d are readily to be obtained using some automatic differentiation software tools such as ADOL-C [14]. Using equations (8) and (9), the total derivative of x…”

Section: A Sensitivity Analysismentioning

confidence: 99%

“…However, empirical studies showed that the generalization performance of such neural network models becomes poor as the dimension of the design space increases [4], [5]. Other related work includes the use of feedforward neural networks to solve ordinary and partial differential equations [6], [7], and discrete-time neural networks [8], or continuous-time (differential) neural network [9] to model dynamic systems. The continuous-time neural networks bring further advantages and computational efficiency over the discrete formulation even if at the end both are represented on the computer using only discrete values [10].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Prediction of convergence dynamics of design performance using differential recurrent neural networks

Cao

Jin

Kowalczykiewicz

et al. 2008

2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence)

View full text Add to dashboard Cite

Abstract-Computational Fluid Dynamics (CFD) simulations have been extensively used in many aerodynamic design optimization problems, such as wing and turbine blade shape design optimization. However, it normally takes very long time to solve such optimization problems due to the heavy computation load involved in CFD simulations, where a number of differential equations are to be solved. Some efforts have been seen using feedforward neural networks to approximate CFD models. However, feedforward neural network models cannot capture well the dynamics of the differential equations. Thus, training data from a large number of different designs are needed to train feedforward neural network models to achieve reliable generalization. In this work, a technique using differential recurrent neural networks has been proposed to predict the performance of candidate designs before the CFD simulation is fully converged. Compared to existing methods based on feedforward neural networks, this approach does not need a large number of previous designs. Case studies show that the proposed method is very promising.

show abstract

“…, d are readily to be obtained using some automatic differentiation software tools such as ADOL-C [14]. Using equations (8) and (9), the total derivative of x…”

Section: A Sensitivity Analysismentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Prediction of convergence dynamics of design performance using differential recurrent neural networks

Cao

Jin

Kowalczykiewicz

et al. 2008

2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence)

View full text Add to dashboard Cite

show abstract

“…Moreover, all the above mentioned dynamic neural models have no state-space description. In fact, approaches trying to solve such a challenging problem can be rarely found in the literature (Pan et al, 2001;Zamarreño and Vega, 1998). Unfortunately, these approaches do not allow calculating the uncertainty of these models, which is necessary to apply them in robust fault detection schemes.…”

Section: Introductionmentioning

confidence: 99%

An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

Mrugalski¹

2013

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

This paper presents an identification method of dynamic systems based on a group method of data handling approach. In particular, a new structure of the dynamic multi-input multi-output neuron in a state-space representation is proposed. Moreover, a new training algorithm of the neural network based on the unscented Kalman filter is presented. The final part of the work contains an illustrative example regarding the application of the proposed approach to robust fault detection of a tunnel furnace.

show abstract

“…Identifying state variables of a complex system is in most cases impossible. However, although artificial neural model states can still correctly describe the impact of system state variable dynamics on the system output (Zamarreno and Pastora, 1998;Kulawski and Brdyś, 2000). This vastly improves the ability of the model to approximate unknown system input-output dynamics.…”

Section: Introductionmentioning

confidence: 99%

Adaptive Prediction of Stock Exchange Indices by State Space Wavelet Networks

Brdyś¹,

Borowa²,

Idźkowiak³

et al. 2009

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

The paper considers the forecasting of the Warsaw Stock Exchange price index WIG20 by applying a state space wavelet network model of the index price. The approach can be applied to the development of tools for predicting changes of other economic indicators, especially stock exchange indices. The paper presents a general state space wavelet network model and the underlying principles. The model is applied to produce one session ahead and five sessions ahead adaptive predictors of the WIG20 index prices. The predictors are validated based on real data records to produce promising results. The state space wavelet network model may also be used as a forecasting tool for a wide range of economic and non-economic indicators, such as goods and row materials prices, electricity/fuel consumption or currency exchange rates.

show abstract

State space neural network. Properties and application

Cited by 75 publications

References 13 publications

Prediction of convergence dynamics of design performance using differential recurrent neural networks

Prediction of convergence dynamics of design performance using differential recurrent neural networks

An unscented Kalman filter in designing dynamic GMDH neural networks for robust fault detection

Adaptive Prediction of Stock Exchange Indices by State Space Wavelet Networks

Contact Info

Product

Resources

About