利用多期均值方法提高新陈代谢GM(1,1)模型的预测精度

Improving the prediction accuracy of metabolic GM(1,1)model by multiphase mean method

在新陈代谢GM(1,1)预测模型的基础上,通过适当选取建模数据维度和增加预测数据长度,使得每期的预测值都有多个新陈代谢GM(1,1)的预测值,最后取多期预测值的平均值作为最终的预测值.数据实验证明了该方法不仅可以进一步提高预测精度,而且能有效消除数据波动,预测数据更具有稳定性,优势明显.

On the basis of metabolic GM(1,1)prediction model,by properly selecting the dimension of modeling data and increasing the length of prediction data,the predicted value of metabolic GM(1,1)in each period has multiple predicted values.Finally,the average value of multi-period predicted value is taken as the final predicted value.Data experiments show that this method can not only further improve the prediction accuracy,but also effectively eliminate data fluctuations.The prediction data is more stable and has obvious advantages.