Study On The Electric Short-term Load Forecasting Based On Distributed Data Mining

This paper presents the load forecasting model of distributed data mining on the basis of load forecasting model by zoning. Firstly, analyzing the framework and characteristic of distributed data mining and its technique support, and applying distributed data mining to the load forecasting. Then, taking BaoDing city and JingJinTang district as instance, the load forecasting system based on distributed data mining is simulated. The influencing factors and the important degree of their and load forecasting model's establishing have been researched, and a method of load forecasting based on RBF neural network combined with decision tree is put forward, which uses RBF neural network to choose the combined effects of forecasting model and evaluate the importance of attributes and uses decision tree algorithm to forecast load. At last, the results demonstrate that the method is feasible.