Chaotic characteristics of electricity price and its forecasting model

In power market environment, electricity price is influenced by many factors and exhibits a very complicated and irregular fluctuation. In order to validate the chaotic characteristic of electricity price, a phase space is firstly reconstructed from the scalar price time series in this paper. Secondly, the main features of attractors, i.e., the correlation dimensions and Lyapunov exponents are extracted and the surrogate data method is used. The analyzed results indicate that electricity price has chaotic characteristic and its short-term forecast can be realized by employing the chaos theory. Then, in order to achieve accurate short-term forecast, in the phase space reconstructed from multivariate time series, the global and local price forecasting model based on the recurrent neural network is proposed and successfully applied to the forecasting of the energy price on the New England market.