An explanation of least squares regression using a linear regression model.

This essay talks about a linear regression model in ordinary least squares regression, the problems that may arise in an OLS model and how it can be fixed through generalized least squares. Generalized Least Squares (GLS) regression is used when problems occur in OLS estimation. The paper includes graphs and formulae supporting the analysis.

“Ordinary Least Squares (OLS) regression is a common tool used for economic forecasting when analyzing time series and cross-sectional data. It is quite common as it is one of the easiest ways of estimating parameters in a simple or multiple regression model. Generalized Least Squares (GLS) regression is used when problems occur in OLS estimation. Such problems will cause are estimators to no longer be the Best Linear Unbiased Estimate (BLUE) or efficient and so it is necessary to use a transformed model. In this essay I will show how both OLS and GLS are derived through equations and discuss the problems that may occur under OLS estimation and why it is more suitable to use GLS to estimate a linear model when certain problems occur.”