A1556
Title: Probabilistic forecasting with machine learning and big data
Authors: Lubos Hanus - UTIA AV CR, v.v.i (Czech Republic) [presenting]
Jozef Barunik - UTIA AV CR vvi (Czech Republic)
Abstract: A distributional deep learning approach is proposed for probabilistic forecasting of economic time series. Being able to learn complex patterns from a large amount of data, deep learning methods are useful for decision making that depends on the uncertainty of a possibly large number of economic outcomes. Such predictions are also informative to decision-makers facing asymmetric dependence of their loss on outcomes from possibly non-Gaussian and non-linear variables. We show the usefulness of the approach on the three distinct problems. First, we use deep learning to construct data-driven macroeconomic fan charts that reflect the information contained by a large number of variables. Second, we obtain uncertainty forecasts of irregular traffic data. Third, we illustrate gains in the prediction of stock return distributions that are heavy-tailed and suffer from low signal-to-noise ratios.
