The problem of missing values of wind power during the collection,transmission and storage brings dif-ficulties to further applications such as wind power prediction based on operating data.Aiming at this problem,a near-neighbor imputation algorithm based on adaptive fuzzy clustering is proposed,named aFCM-KNN.In view of the strong randomness and volatility of the wind power data itself,the wind power data is clustered based on the FCM al-gorithm according to the wind speed.In order to solve the problem that the number of clusters needs to be manually set by FCM,an adaptive strategy for determining the number of clusters is designed according to the distribution char-acteristics of wind power data.Considering that the direct imputation after clustering is susceptible to noise,the KNN algorithm is used to impute the missing values in each subcluster according to the neighbor of the sample,which fur-ther improves the imputation accuracy.Test results on actual data show that the method has a higher accuracy than the other six commonly used imputation algorithms.
关键词
风电功率/缺失值填补/模糊均值聚类/近邻算法
Key words
Wind power/Missing value imputation/Fuzzy means clustering/Nearest neighbor
维普
万方数据