What is Regression?
Regression is the mathematical expression of the relationship between variables – dependent and independent – in such a functional form that the dependent variable is said to be regressed on the independent variable(s). Eg, Regression Y on X or Y = f(X).
Regression is simple when only two variable are in question.
The purposes of regression analyses are a combination of the following;
- Testing Theory – the purpose of research is analysis.
- For forecasting and prediction.
- For policy recommendations.
Based on the Ordinary Least Squares Technique (OLS), regression seeks to minimize the square of the negative and positive deviations resulting from the line of best for that describes the mathematical, statistical and econometric relationship between two variables. It has the prime “BLUE – Best Linear Unbiased Estimate” ; a quality not present in other estimators.