Study of analytical function of correlation for random parameter process

Abstract

In this article, we are presenting one modeling method to obtain analytic functions, corresponding to the statistical correlation of parameters of an industrial process, where some parameters of the process are influenced by permanent variables and the entire process is also influenced by random events. The proposed method of this study, involves finding the polynomial functions for real distribution of probability, for the individual set of variables, according to the histogram of relative frequency. Those functions have the property that generated surfaces of the function from definition domain to be unitary - this will be the normalization condition. By eliminating the common parameter of functions, we obtain a correspondence relationship, between that functions, which we will call analytical correlation function of the two random variables. This method is a numerical solution of the problem given and applied successfully in statistical analysis of multi-parametric industrial processes, where he proved fully effective.