Beta autoregressive moving average model selection with application to modeling and forecasting stored hydroelectric energy

We evaluate the accuracy of model selection and associated short-run forecasts using beta autoregressive moving average (,KARMA) models, which are tailored for modeling and forecasting time series that assume values in the standard unit interval, (0, 1), such as rates, proportions, and concentration indices. Different model selection strategies are considered, including one that uses data resampling. Simulation evidence on the fre-quency of correct model selection favors the bootstrap-based approach. Model selection based on information criteria outperforms that based on forecasting accuracy measures. A forecasting analysis of the proportion of stored hydroelectric energy in South Brazil is presented and discussed. The empirical evidence shows that model selection based on data resampling typically leads to more accurate out-of-sample forecasts. (c) 2021 International Institute of Forecasters. Published by Elsevier B.V. All rights reserved.