This paper is concerned with the forecasting performance of variable reduction and variable selection methods using medium and large datasets. The variable reduction methods include Principal Components, Partial Least Squares and Bayesian Shrinkage Regression. The variable selection methods...