Error Correction Model for Short-term Electricity Price Forecasting

Based on the theory of time series error correction model,this paper attempts to establish the error correction model(ECM) to support the short-term electricity price forecasting.The data of electricity price and its influencing factors are preprocessed for stationarity through the difference.The autoregressive exogenous(ARX) model is established and the co-integration relationship between the two variables is checked by augmented dickey-fuller(ADF) method.Therefore,the ECM for forecasting electricity price can be built up by the evaluated model parameters.Compared with the traditional time series model,the ECM has merits in effectively modeling the short-term fluctuations and contributing to forecasting the trend of electricity price.The model takes the main influencing factors of electricity price into account,with focus on analyzing the inherent change rules of electricity price.The forecasting results of PJM in the USA show that the proposed model can offer a higher forecast precision than the traditional time-series methods.