摘要
针对GM(1,1)模型对波动序列进行模拟、预测时通常存在较大误差的问题,提出用时间系数对等间距时序进行修正,给出了计算时间系数的方法;根据时间系数的特点利用反向累加生成的GOM(1,1)模型,建立GM(1,1)模型与GOM(1,1)模型相结合的两阶段灰色模型,进一步拓展了灰色模型的适用范围.结果表明,提出的两阶段灰色模型能够适应于有较大波动的原始数据序列的分析和建模,且具有一定的实用性与可靠性.
As there's big simulation and forecast error when modeling toward fluctuant data sequence with GM(1,1), this paper proposes to amend the equal interval time sequence with time coefficient, and gives the computing method. According to the characteristic of time coefficient and the accumulated generating operation in opposite direction in grey model GOM(1,1) ,we put forward the two stage grey model by combining GM(1,1) model with GOM (1,1) model, and therefore broaden the applicable scope of grey model. The example indicates that the two stage grey model has the superiority in modeling toward the big fluctuant data sequence, moreover has certain usability and reliability.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2008年第11期109-114,121,共7页
Systems Engineering-Theory & Practice
基金
国家自然科学基金(70473037)
江苏省软科学重点项目(BK2006025)