Modeling and predicting U.S. recessions using machine learning techniques

The most representative machine learning techniques are implemented for modeling and forecasting U.S. economic activity recessions in particular. An elaborate, comprehensive, comparative framework is employed order to estimate recession probabilities. empirical analysis explores the predictive content of numerous well-followed macroeconomic financial indicators, but also introduces a set less-studied predictors. ability underlying models evaluated using plethora statistical evaluation metrics. results strongly support application over more standard econometric area prediction. Specifically, indicates that penalized Logit regression models, k-nearest neighbors, Bayesian generalized linear largely outperform ‘original’ Logit/Probit prediction recessions, as they achieve higher accuracy across long-, medium-, short-term forecast horizons.