The estimation of the total standard and extended uncertainty for the neural network model of nonlinear measuring transducer based on three-layer perceptron is considered. The source of the uncertainty is neural network synaptic weights uncertainty. The uncertainty behavior of the neural network model is researched by the computer simulation, the comparative analysis with polynomial model uncertainty is implemented.
transfer function, artificial neural network, perceptron, model, uncertainty
"Otsenyvanye neopredelennosty neirosetevoi modely nelyneinoho yzmerytelnoho preobrazovatelia" [The uncertainty estimation for the neural network model of the nonlinear measuring transducer],
Information Processing Systems,