“…This approach can be generalized by assuming that the finding is not subject to a discrete set of evaluation results or predictions, and a sequence of functions of a continuous argument, defined by a sequence of finite intervals, which are hereinafter referred to intervals of analysis. These functions include, in particular, is a generalization of polynomials (polynomials of a given system of basis functions) or the regression function obtained in the course of solving a sequence of least squares problems [3,4]. If, moreover, requires that the local evaluation results were represented generally smooth function that necessitates a smooth interfacing of said polynomials, the least squares problem is converted into a quadratic optimization problem with constraints in the general case, as the type of equality and inequality [4 -6].…”