Short-term forecasting of electricity prices with a computationally efficient hybrid approach

Electricity price forecasting models are of great importance for market participants due to their considerable volatility, especially in deregulated and competitive contexts. As a result, these models are highly demanded, especially in day-today applications, which require not only accurate results, but also fast responsiveness. Taking these needs into account, this work proposes a novel short-term electricity forecasting approach by means of a hybrid model, combining econometric and fundamental methods. In order to validate this work's proposed method under complex price dynamics, the model has been tested for the Iberian electricity market case, and further verified by comparing its performance with other, more traditional, forecasting models.

I. Introduction

Since the deregulation and liberalisation of electricity markets, its participants and agents are increasingly being exposed to price uncertainty and high market volatility. Moreover, the growing penetration of renewable energy sources and global financial instability adds even more variability to the market, which strongly encourages the use of electricity price forecasting models.