Generalized Least-Squares Regressions V: Multiple Variables

Authors

Nataniel Greene, CUNY Kingsborough Community CollegeFollow

Document Type

Article

Publication Date

3-2015

Abstract

The multivariate theory of generalized least-squares is formulated here using the notion of generalized means. The multivariate generalized least-squares problem seeks an m dimensional hyperplane which minimizes the average generalized mean of the square deviations between the data and the hyperplane in m + 1 variables. The numerical examples presented suggest that a multivariate generalized least-squares method can be preferable to ordinary least-squares especially in situations where the data are ill- conditioned.

Comments

This work was originally published in New Developments in Pure and Applied Mathematic.