A Hybrid Model for Forecasting Short-Term Electricity Demand

TL;DR

This paper introduces HYENA, a hybrid model combining feature engineering, window predictors, and LSTM encoder-decoders, achieving state-of-the-art accuracy in short-term UK electricity demand forecasting.

Contribution

The paper presents HYENA, a novel hybrid forecasting model that significantly improves accuracy over existing models for UK electricity demand prediction.

Findings

HYENA reduced MAPE by 16%.

HYENA decreased RMSE by 10%.

HYENA sets new state-of-the-art in UK load forecasting.

Abstract

Currently the UK Electric market is guided by load (demand) forecasts published every thirty minutes by the regulator. A key factor in predicting demand is weather conditions, with forecasts published every hour. We present HYENA: a hybrid predictive model that combines feature engineering (selection of the candidate predictor features), mobile-window predictors and finally LSTM encoder-decoders to achieve higher accuracy with respect to mainstream models from the literature. HYENA decreased MAPE loss by 16\% and RMSE loss by 10\% over the best available benchmark model, thus establishing a new state of the art for the UK electric load (and price) forecasting.