CALCULATION OF PARETO-OPTIMAL ESTIMATES OF REGRESSION PARAMETERS

The parameters of a linear regression in the presence of multicollinearity of the regressors are estimated. A generalization of the comb estimate is proposed by introducing an n-dimensional unknown vector (n is the number of regression parameters estimated). The sought estimate is found in two stages: at the first stage, the vector mentioned above is determined. and at the second stage, the regularization parameter used in the usual comb regression is found. Calculations by the method proposed are illustrated by examples.