Analytical dependences between the determination coefficients and the ratio of error variances of the test items in Deming regression model

The article deals with the problem of building regression models, in which all variables have stochastic nature. To solve it, we propose to use the determination coefficient. We obtain analytical dependencies of the determination coefficients from the ratio of error variances of the test items. We set the optimization problem, assuming the maximization of the determination coefficients sum for each Deming regression equation. We give a model example of the numerical processing of Deming regression with its parameters and sign errors which are known.