Answers are grounded in this paper and cite their sources.
Authors demonstrate a high-accuracy output estimate in a linear stochastic system, suggesting a novel application of wavelet neural networks.
Paper is devoted to synthesis by integral functional Bayes (FBC) criterion of one dimensional linear stochastic system with high availability (StSHA) based on wavelet neural network (WNN) and canonical expansions (CE) of stochastic process (StP). StSHA is described by Pugachev equations. Input and output signals are linear for random parameters. Input contains Gaussian additive noise. Random parameters do not depend on random noise. FBC optimal estimate of output is constructed using CE StP. For output estimate multi-layer WNN is constructed. Optimal values of unknown coefficients are calculated by WNN learning. Accuracy estimation is performed by mathematical expectation and variance of required output signal. Test example demonstrates high accuracy of method and algorithm. Experimental software tools for informational StSHA quality analysis and synthesis are developed. Perspectives for future investigation are outlinear.
No takes yet. Share an insight, caveat, or question.
Sinitsyn et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: