Electricity Demand Forecasting by Multi-Task Learning

TL;DR

This paper applies kernel-based multi-task learning to forecast electricity demand across multiple nodes, effectively modeling complex seasonal patterns and inter-node correlations, leading to improved predictive accuracy.

Contribution

It introduces the use of output kernel learning techniques with multiplicative structures for electricity demand forecasting, outperforming traditional additive models.

Findings

Kernel methods effectively model complex seasonal demand patterns.

Multiplicative kernels outperform additive models in prediction accuracy.

Multi-task learning leverages correlations between demand profiles.

Abstract

We explore the application of kernel-based multi-task learning techniques to forecast the demand of electricity in multiple nodes of a distribution network. We show that recently developed output kernel learning techniques are particularly well suited to solve this problem, as they allow to flexibly model the complex seasonal effects that characterize electricity demand data, while learning and exploiting correlations between multiple demand profiles. We also demonstrate that kernels with a multiplicative structure yield superior predictive performance with respect to the widely adopted (generalized) additive models. Our study is based on residential and industrial smart meter data provided by the Irish Commission for Energy Regulation (CER).