Short-term Load Forecasting of Power Model Based on CS-Catboost Algorithm

Accurate prediction of short-term power load is of great significance to the planning of distribution network and the safety and reliability of power system. In order to further improve the accuracy of short-term load prediction of power system, this paper proposes a power load prediction model based on CS-Catboost algorithm. Utilize the global search ability of the Cuckoo Search algorithm to optimize the hyperparameters in the Catboost algorithm, thereby increasing the accuracy of the model prediction. In this paper, several regions in Spain are used as examples to conduct experiments, considering various factors such as weather and time, to predict the short-term power load in one day. The results show that compared with the unoptimized Catboost algorithm, LightGBM algorithm, GBDT algorithm and RF algorithm, the MSE and MAPE of CS-Catboost algorithm prediction results are reduced by 2027.43MW, 3192MW, 3619MW, 8327MW and 0.14%, 0.17%, 0.29%, 0.57%, and the prediction is more accurate at the peak and valley of the power load, and has high practical value.