Application of Harmonic Clustering and Classification Method in Electric Power Load Forecasting

Clustering and classification are two important research areas of data mining, and classification needs prior-knowledge, while clustering is often based on a similar measure to find its own inherent characteristics from the data. In practice, the results of classification and clustering are often inconsistent. For this problem, the definition of harmonic matrix is given in this paper, and based on this conception a harmonic clustering-classification algorithm is proposed, which makes the results of classification and clustering keep high consistency. At the same time, this method has been used in power system load forecasting, and the experiment shows that the classification results obtained by our method are more reliable.