The Combination Forecasting of Electricity Price Based on Price Spikes Processing: A Case Study in South Australia

Electricity price forecasting holds very important position in the electricity market. Inaccurate price forecasting may cause energy waste and management chaos in the electricity market. However, electricity price forecasting has always been regarded as one of the largest challenges in the electricity market because it shows high volatility, which makes electricity price forecasting difficult. This paper proposes the use of artificial intelligence optimization combination forecasting models based on preprocessing data, called “chaos particles optimization (CPSO) weight‐determined combination models.” These models allow for the weight of the combined model to take values of [−1,1]. In the proposed models, the density‐based spatial clustering of applications with noise (DBSCAN) algorithm is used to identify outliers, and the outliers are replaced by a new data‐produced linear interpolation function. The proposed CPSO weight‐determined combination models are then used to forecast the projected future electricity price. In this case study, the electricity price data of South Australia are simulated. The results indicate that, while the weight of the combined model takes values of [−1,1], the proposed combination model can always provide adaptive, reliable, and comparatively accurate forecast results in comparison to traditional combination models.