摘要
GM(1,1)具有要求样本数据少,运算方便,短期预测精度高等优点,因此而得到广泛应用;但它也存在一定的局限性,其主要适用于单一的指数增长的模型,对序列数据出现异常情况很难预测。为了改善这一局限性,将其与多项式拟合进行组合,通过实例分析证明组合模型比单一模型拟合精度更高,达到比较好的效果。
GM (1, 1) has the sample data requirements, convenient operation, less short-term prediction accuracy. So it has widely application. But, it also has some limitations , such as it main apply to single exponential growth model. In order to improve the limitations, we mixed the apolynomial fitting portfolio and GM (1,1) in this paper . The example analysis proved that the fitting accuracy of Combination model was higher than the single model portfolio model.
出处
《地理空间信息》
2012年第4期140-141,183,共2页
Geospatial Information
基金
国家自然科学基金资助项目(41071294)
广西科学基金资助项目(桂科基0991023)
广西区研究生科研创新资助项目(2010105960816M19)
广西研究生教育创新计划资助项目(2011105960816M13)
