Electricity consumption forecasting in peak load month based on variable weight combination forecasting model

According to the good growth characteristics of GM (1, 1), the growing trend of the monthly load time series is simulated with vertical historical load data as samples. According to the characteristics of ARIMA model that can better describe the non-stationary data series, the growing trend of the monthly load is simulated with horizontal historical load data as samples. The variable weight is introduced, and the variable weight combination forecasting model combining the merits of GM (1, 1) and ARIMA model is established, which is then applied to forecast the electricity consumption in the peak load month. Experiment results compared with single forecasting model show that the method has a more stable and less forecasting error.