Daily middle-term probabilistic forecasting of power consumption in North-East England

Probabilistic forecasting of power consumption in a middle-term horizon (few months to a year) is a main challenge in the energy sector. It plays a key role in planning future generation plants and transmission grid. This paper proposes a novel model that (i) incorporates seasonality and autoregressive features in a traditional time-series analysis and (ii) includes weather conditions in a parsimonious machine learning approach, known as Gaussian Process. Applying to a daily power consumption dataset in North East England, provided by one of the largest energy suppliers, we obtain promising results in Out-of-Sample density forecasts up to one year, even using a small dataset, with only a two-year calibration set. For the evaluation of the achieved probabilistic forecasts, we consider the pinball loss-a metric common in the energy sector-and we assess the coverage-a procedure standard in the banking sector after the introduction of Basel II Accords-also running the conditional and unconditional tests for probability intervals. Results show that the proposed model outperforms benchmarks in terms of both accuracy and reliability.