基于多因素指数-灰色GM(1,1)模型的能耗分析与预测

Energy Consumption Analysis and Prediction Based on Multifactor Index-GM(1,1)Model

针对烟草企业能耗数据波动大且传统预测模型预测效果不佳的问题,提出了多因素指数-灰色GM(1,1)模型.首先,梳理影响能耗波动的外部因素,结合相关系数获得各外部影响因素的相关性排序.其次,利用逐步回归分析法建立多元线性回归模型,计算多因素指数以平滑能耗序列.最后,对平滑后的序列分组并建立非等时距模型,还原预测值得到最终结果.实验分析表明,提出的多因素指数-灰色GM(1,1)模型可以显著提高烟草企业能耗数据的预测精度.

In order to solve the fluctuation of tobacco enterprises energy consumption and the poor predictive result of the traditional model,a multifactor index-GM(1,1)model is proposed.Firstly,the external factors that affect the fluctuation of energy consumption are sorted out,and the correlation of all external factors is judged by the correlation coefficient.Secondly,stepwise regression analysis is used to establish a multiple linear regression model of external factors.Then a multiple factor index is obtained to smooth the original data.Finally,the smoothed sequences are further grouped and a non-isochronous model is established,and the final prediction results are obtained by restoring the predicted values.Experimental analysis shows that the multifactor index-GM(1,1)model proposed in this paper can significantly improve the prediction accuracy.