Prediction Error and Forecasting Interval Analysis of Decision Trees with an Application in Renewable Energy Supply Forecasting

Abstract

Renewable energy has become popular compared with traditional energy like coal. The relative demand for renewable energy compared to traditional energy is an important index to determine the energy supply structure. Forecasting the relative demand index has become quite essential. Data mining methods like decision trees are quite effective in such time series forecasting, but theory behind them is rarely discussed in research. In this paper, some theories are explored about decision trees including the behavior of bias, variance, and squared prediction error using trees and the prediction interval analysis. After that, real UK grid data are used in interval forecasting application. In the renewable energy ratio forecasting application, the ratio of renewable energy supply over that of traditional energy can be dynamically forecasted with an interval coverage accuracy higher than 80% and a small width around 22, which is similar to its standard deviation.

Forecasting automobile gasoline demand in Australia using machine learning-based regression

“…Machine learning, proposed by Samuel (1959), has been widely used in the fields of energy and economics with diverse applications, for example, the optimization of energy inputs Abdelaziz et al, 2016;Nabavi-Pelesaraei et al, 2017;Ali & Abd Elazim, 2018;Khanali et al, 2021), the investigation of energy efficiency , forecasting energy commodity prices (Ding, 2018;Yu et al, 2017;Zhang et al, 2015), forecasting energy demand (Yang et al, 2014;Panapakidis and Dagoumas, 2017;Ou et al, 2020;Haque et al, 2021). Popular ML techniques in the relevant literature include applied artificial neural networks (ANN) (Olanrewaju et al, 2013;Kunwar et al, 2013); deep learning (Lago et al, 2018;Peng et al, 2018), support vector machine (SVM) (Papadimitriou et al, 2014;Zhu et al, 2016;Jiang et al, 2018), decision trees (Bastardie et al, 2013;Zhao and Nie, 2020) and ensemble methods (Ghasemi et al, 2016;Mirakyan et al, 2017).…”