With the flourishing development of the electric undertaking in recent years, the requirement forecasting of the electricity seems particularly important. This text regards a large number of regional historical electric consumption data as the foundation of analyzing, through predicting to more than ten kinds of commonly used main medium and long-term electricity at present the prediction data error rate of the method is compared, has synthesized the quintessence of various prediction methods creatively, and according to the expert's knowledge and experience in the field, select to use familiar to all simple EXCEL apt to use as the developing platform, by way of deducing the bold comprehensive∑'s weighting of adoption, predict models after putting forward∑'s weighting of comprehensive analysis. Result of study indicate, use system this carry on error rate that load predict 4% about only, and has reduced development of the software and application cost, has shortened cycle of predicting greatly, can avoid the appearance of the artificial convoluted and artificial mistake that deduce at the same time, make the best result of predicting fast. |