Effect of Multicollinearity on Variable Selection in Multiple Regression

When Multicollinearity exists in a data set, the is considered deficient. frequently encountered observational studies. It creates difficulties when building regression models. phenomenon whereby two or more explanatory variable multiple model are highly correlated. Variable selection an important aspect of as such choice best subset among many variables to be included most difficult part analysis. Data was obtained from Nigerian Stock Exchange Fact Book, Annual Report and Account, CBN Statistical Bulletin FOS bulletin 1987 2018. Variance Inflation Factor (VIF) correlation matrices were used detect presence multicollinearity. Ridge Least Square Regression applied using R-package, Minitab SPSS Packages. Models with constant range 0.01 ≤ K 1.5 models for each value P = 2, 3, …,7. The optimal by taking average rank Coefficient Determination Mean Error. result showed that choices affected multicollinearity different selected under same level P.