Research on Annual Electric Power Consumption Forecasting Based on Partial Least-Squares Regression

With the deterioration of primary energy market supply, it is important to optimize the raw material buying and dispatching. The annual electric power consumption is one of the most important decision making basis to realize this. Because of the characters of observations, OLS method and neural network model are all not suit for this. PLS extract variables one by one from few historical data. Under the control of modeling, it makes fully use of the useful information contained in the raw data. The experiments show that this method is feasible in annual electric power consumption forecasting.