|
on Forecasting |
| By: | Jozef Barunik; Lubos Hanus |
| Abstract: | We propose a novel machine learning approach to probabilistic forecasting of hourly intraday electricity prices. In contrast to recent advances in data-rich probabilistic forecasting that approximate the distributions with some features such as moments, our method is non-parametric and selects the best distribution from all possible empirical distributions learned from the data. The model we propose is a multiple output neural network with a monotonicity adjusting penalty. Such a distributional neural network can learn complex patterns in electricity prices from data-rich environments and it outperforms state-of-the-art benchmarks. |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:arx:papers:2310.02867&r=for |
| By: | Erick Inácio Ferreira (UFMG); Igor Viveiros Melo Souza (UFMG) |
| Abstract: | The aim of this study is to assess the performance of two well-known algorithms which automate the process of modeling and forecasting time series, each applying a different econometric technic: ARIMA or exponential smoothing. We provide a brief discussion of how these algorithms work and results of a Monte Carlo experiment, which was conducted to evaluate the capabilities of auto.arima and ets, available in Rob Hyndman’s forecast package for the statistical software R, commonly used by economists to study and forecast time series. Over 200.000 synthetic series were simulated, with several different characteristics, used to test both methods and report metrics of correct modeling and out-of-sample forecast errors of the algorithms, on top of which we provide a brief discussion of the successes and shortcomings that happened while applying each algorithm. |
| Keywords: | Time series econometrics, ARIMA, exponential smoothing, auto.arima |
| Date: | 2023–10 |
| URL: | http://d.repec.org/n?u=RePEc:cdp:texdis:td661&r=for |