论文部分内容阅读
针对传统灰色GM(1,1)预测模型维数确定困难、适用范围小和预测精度不高等局限性,提出了一种能处理复杂序列的动态的最佳维数GM(1,1)幂模型.最后以2003-2013年居民收入基尼系数为研究样本做预测分析,同时建立了传统GM(1,1)模型、经典GM(1,1)幂模型作为对比,结果表明:动态的最佳维数GM(1,1)幂模型的平均相对误差为0.08%,显著低于传统GM(1,1)模型的1.04%和经典GM(1,1)幂模型的0.85%.
In order to overcome the limitations of the traditional gray GM (1,1) prediction model, such as difficulty of dimension determination, small scope of application and low prediction accuracy, a dynamic optimal dimension GM (1,1) power model capable of handling complex sequence is proposed. Finally, based on the Gini coefficient of residents’ income from 2003 to 2013, the forecasting analysis is made and the traditional GM (1,1) model and the classic GM (1,1) power model are contrasted. The results show that the dynamic best dimension The average relative error of the GM (1,1) power model is 0.08%, which is significantly lower than 1.04% of the traditional GM (1,1) model and 0.85% of the classical GM (1,1) power model.