Generalized Least Square Method for Regression Analysis

Abstract

In regression analysis while estimating the parameters our data should follow the assumption of classical regression to have the valid estimator of the regression parameters. Sometime while estimating the parameter data does not follow the assumption of classical regression like constant variance of residual and autocorrelation of residual in that situation classical approach of estimating the regression parameter ordinary least square will not give valid estimators in that situation generalized least square will used.