Parametric and nonparametric estimation for characteristics of randomized models under limited data (entropy approach)

The paper presents a new approach to determination of dependencies between limited input and output data. It is based on randomized static and dynamic models and on estimation of the probabilistic characteristics of their parameters. The static and dynamic models are described by the functional polynomials. Entropy approach is developed for parametric and nonparametric estimation, for which generalized informational Boltzmann's and Fermi's entropies are used.