The model combination method of power system load forecasting based on freshness availability index

This paper deeply analyzed the merit and demerit of current screening techniques of power system load forecasting models. Taking the time-based data weight into consideration, a freshness function is introduced to define the freshness availability index as an evaluation criterion for load forecasting models. The defined freshness availability index is also used for constructing screening technique of load forecasting model and the variable weight combination forecast method of medium and long term power system load forecasting. The case studies show a highly precise forecasting result and good practicability of the proposed combination forecast method.