Short-Term Wind Power Prediction Method Based on Combination of Meteorological Features and CatBoost

As one of the hot topics in field new energy, short-term wind power prediction research should pay attention to impact meteorological characteristics on while improving accuracy. Therefore, a method based combination features and CatBoost is presented. Firstly, morgan-stone algebras sure independence screening(MS-SIS) designed filter features, influence explored. Then, sort enhancement algorithm increase accuracy calculation efficiency reduce risk single element. Finally, network constructed further realize prediction. The National Renewable Energy Laboratory (NREL) dataset used for experimental analysis. results show that not only improve power, but also have higher efficiency.