GRADING PARAMETERS OF A REGRESSION MODEL BASED ON SINGULAR VALUE DECOMPOSITION MATRICES
Abstract
The problem of estimating the parameters of unconditional linear regression models in conditions of poor conditionality of the correlation matrix, quasimulticollinearity, cannot be precisely solved when using the classical least squares method (LSM). Developed a method for estimating the parameters of a regression model based on the LSM and the singular value decomposition of matrices of empirical data in case of quasimulticollinearity and poor conditionality of the correlation matrix