Regularization methods for the short-term forecasting of the Italian electric load

TL;DR

This paper introduces regularization techniques for short-term Italian electric load forecasting, framing it as a multitask learning problem and demonstrating improved accuracy over existing methods, with potential for further enhancement through forecast aggregation.

Contribution

It proposes novel regularization approaches for multitask learning in load forecasting, effectively reducing model complexity and improving prediction accuracy compared to Terna.

Findings

Outperformed Terna in quarter-hourly mean absolute percentage error

Forecast aggregation further reduced errors by up to 30%

Regularization methods effectively controlled model complexity

Abstract

The problem of forecasting the whole 24 profile of the Italian electric load is addressed as a multitask learning problem, whose complexity is kept under control via alternative regularization methods. In view of the quarter-hourly samplings, 96 predictors are used, each of which linearly depends on 96 regressors. The 96x96 matrix weights form a 96x96 matrix, that can be seen and displayed as a surface sampled on a square domain. Different regularization and sparsity approaches to reduce the degrees of freedom of the surface were explored, comparing the obtained forecasts with those of the Italian Transmission System Operator Terna. Besides outperforming Terna in terms of quarter-hourly mean absolute percentage error and mean absolute error, the prediction residuals turned out to be weakly correlated with Terna, which suggests that further improvement could ensue from forecasts aggregation. In fact, the aggregated forecasts yielded further relevant drops in terms of quarter-hourly and daily mean absolute percentage error, mean absolute error and root mean square error (up to 30%) over the three test years considered.