Load Forecasting of Green Power System Based on Data Mining

This project applies Data Mining technology to the prediction of electric power system load forecast. It proposes a mining program of electric power load forecasting data based on the similarity of time series research, in the method, the average of each segment of time-sequence load is used to reduce the dimension of the problem. The similarity inquiring of each sub-sequence of loads is realized by using slipping window and MBR method. The inquiring is improved in efficiency by designing the index structure according to the-tree. Effectively overcome the negative effects on the prediction results caused by the limited and incomplete data. It also illustrates a list of examples to prove that the conducted method is effective and efficient.