Forecasting with many predictors: an empirical comparison

Three methodologies of estimation of models with many predictors are implemented to forecast Colombian inflation. Two factor models, based on principal components, and partial least squares, as well as a Bayesian regression, known as Ridge regression are