Sparse electricity load forecasting based on ARIMA and random forest models

Accurate prediction of power load is the basis of power dispatching, which plays an important role in the security, stability and economy of power system operation, and it is necessary to carry out accurate prediction of power load. Aiming at the sparse daily power load data, this paper uses the random forest model on the basis of ARIMA model to forecast the power load of 220 daily power load data in 2020, and conducts performance tests. Not only the prediction results are compared and analysed with real data, but also other models are compared horizontally. The results show that the combination of ARIMA model and random forest has excellent prediction effect, and the combination of random forest model and ARIMA model is more suitable for daily electricity load prediction than the long and short-term memory network model.