A Short-Term Wind Speed Prediction Method Based on Optimal Copula Correlation Analysis
Accurately analyzing the correlations among variables and deeply exploring the potential value of the data are the key to improving the accuracy of a series of wind speed prediction models constructed based on the principles of statistical analysis.To maximize the potential value of data preservation and eliminate redundant information,the input variables to be selected are firstly fitted with probability density functions,then the optimal Copula functions are constructed among wind speed and other variables,furthermore the correlation coefficients are solved in accordance with the optimal Copula functions to clarify the key input variables affecting the accuracy of the wind speed prediction,and finally the prediction results are obtained based on long short-term memory network model.The proposed method is validated based on the measured dataset of a certain region in China.The experimental results show that the proposed method could effectively select the key input variables and improve the prediction accuracy while reducing the model training time.
wind speed predictioninput variable selectioncorrelation analysisCopula function
NETL
NSTL
万方数据
