DOI: 10.36347/sjebm.2021.v08i01.003

Pages: 14-25

The output of four variable selection techniques in the construction of a model that best estimates a dependent variable is critically evaluated in this analysis. The techniques for variable selection are the direct search on the t equation, the method of forward selection, the method of backward exclusion, and the method of stepwise regression. In contrast, economic data of 32 years were collected on Real Gross Domestic Product each, that was the dependent variable used as a measure of economic development and growth, and seven factors; Growth Market Capitalization, All-Shares Index, Market Turn-Over, Nigerian Trade Economy Transparency, Transaction Value, Nigerian Stock Exchange Total Listing. Using the four variable selection techniques, the residual mean square, modified R^2, and the variance inflation factor obtained from the use of each of these techniques, which are the criteria for determining the best model, the actual gross domestic product was compared with the seven variables by rating them on the basis of the formula that best met the evaluation criteria. The result shows that the backward elimination method performs better in variable selection based on the sample collected with a mean rank of 1.67 taken across the parameters, and supports the use of the all possible combination method as a control.