离散DGM(1,1,t~α)模型

Discrete DGM(1,1,t~α) Model

针对具有部分指数特征且含时间幂次项的数据序列,分别建立了以x^0(1)和x^1(n)为初始条件的离散DGM(1,1,t~α)模型,并给出了α取1,2时的模型公式及其推导过程;根据数据序列的最小二乘曲线并不一定就是由初始值迭代拟合生成的曲线,对DGM(1,1,t~α)模型的迭代基值进行了优化,并给出了基值修正项的估计式;最后,通过一个实例验证了所建模型的合理性与实用性.

In view of the data sequences who contain partial index characteristics and time exponential,the discrete DGM（l,l,t~α） model was established respectively by x^0（1） and the x^1（n） as the initial condition,the specific derivation process of the model were also given whenα is equal to 1,2.Based on the least squares of the data sequence is not necessarily hopson into the curve of the plans drawn up by the initial value iteration,the iterative base value of the DGM（l,l,t~α） model is optimized,and the base value estimation formulas of correction term are also given.Finally,the rationality and practicality of the model is verified by an example.