A Novel Data-Driven Method with Decomposition Mechanism Suitable for Different Periods of Electrical Load Forecasting

For improving the precision of load forecasting in different time spans, a new model which combines improved complete ensemble empirical mode decomposition algorithm based on adaptive noise (ICEEMDAN) algorithm, least squares support vector machine (LS-SVM) and long short-term memory network (LSTM) is proposed. In this paper, training set acquired from department dispatch center large city north China. And advantages algorithms are applied reasonably, where ICEEMDAN used to decompose original historical data. By using fluctuation trend periods can be obtained. structure neural combining LS-SVM LSTM, obtain result. Based LSTM non-stationary stationary signals have been processed, respectively. The simulation results show that whether testing by dataset China or Elia dataset, proposed method outperforms short-, medium- long-term PSO-SVR, LS-SVM, ICCEMDAN-LSTM,