Application of Data Mining Technologies for Forecasting Individual Load

Abstract

This paper proposes a new individual load forecasting method based on data mining technologies. First, the hierarchical cluster analysis is performed to analyze the historical daily load. Next, the results of hierarchical clustering, the week type and weather factors are used to establish a decision tree. Then, the SVM model is established on history days which are similar with the forecast date. Finally, the forecasted load is obtained through the SVM forecasting model. Case studies using real load data show that the proposed new method can improve the accuracy of individual load forecasting.