Reinforcement Learning : An Artificial Intelligence Approach to Forecasting when There Are Structural Breaks

Economic forecasting during structural breaks is challenging due to the possible systematic failure of existent models. Robust forecast devices are able to provide unbiased forecasts just after structural change but at the cost of higher variance in normal times. Reinforcement learning (RL) algorithms can help to discover better policies in environments with little or no information on about the environment and adapt quickly to them. We provide evidence based on real data for inflation and Monte Carlo exercises that these types of algorithm can improve forecast performance in a scenario of structural change without loosing significant performance under normal times