As for the particularity of the power load forecasting,
this paper presents a new power load combination forecasting
model, that is, first select several basic load forecasting methods
to forecast the results, and then select an improved particle
swarm optimization (PSO) algorithm to optimize the weights of
single-forecast results. According to the weights, calculate the
weighted average of the single-forecast results, and the obtained
value is the load forecast value. The empirical results show that
the combination forecasting model combines the advantages of
various methods, and greatly improves the accuracy of power
load forecasting.