The prediction of wind power interval based on k-means clustering algorithm and hoeffding tree classification algorithm