A study on the structure of asymmetric volatility in load series

With the development of power system and the process of electricity market, load forecasting is paid more and more concern to nowadays. In a sense, the precision of load forecasting is crucial to the operation and dispatch of power system while further discussions on the characteristics of the load time series are beneficial. In this paper, by using the rolling windows approach, EGARCH, EGARCH-t and EGARCH-GED model series are estimated. The analysis on the dynamic significance line (DSL) indicates the stability of the reverse leverage effect (RLE) over all the sub-sample space. The analysis of the responses in volatility of EGARCH model and the rediscovery of asymmetry characteristic are provided by employing the news impact curve (NIC). The conclusion is enhanced with the help of PARCH model series also obtained by rolling windows approach. Finally, the forecasting performance is validated by a comparison. This theoretical discussion of asymmetry will help to understand the inherent characteristics of the load time series.