On the paradox of models with a random frequency possessing a given correlation function and standard one-dimensional distribution density
A A Soloveychik · Tractors and Agricultural Machinery · 2012
Models of stochastic processes possessing the given correlation function and the standard one-dimensional distribution density are analyzed. It is shown that contrary to expectations the introduction of auxiliary condition of one-dimensional distribution density normality into a random frequency model substantially worsens the reproduction accuracy of Gaussian processes.