摘要
根据模型的指数特性以及积分特点,以数据曲线在每个区间[k_(i-1),k_i]与坐标轴所围成的梯形面积作为模型背景值z^(1)(k_i)与z^(1)(k_(i-1))的差值,并对其进行修正,从而达到对传统非等间距GM(1,1)模型进行优化的目的.最后采用实例进行验证,并将结果同其他文献的拟合精度进行对比,从而验证算法的有效性与可行性.
According to the exponential properties and intergral characteristics of model,used the trapezoidal area surrounded by the interval [k(i-1),ki] and the axis as the difference of background value z^(1)( ki) and z^(1)(k(i-1)),then revised it. Thus achieved the goal to optimize the traditional non- equidistant GM( 1,1) model. We tested the model with real data,and compared the fitting precision with other literatures,the results showed the model's feasibility and efficiency in this paper.
出处
《福州大学学报(自然科学版)》
CAS
北大核心
2016年第3期306-310,314,共6页
Journal of Fuzhou University(Natural Science Edition)
基金
国家自然科学基金资助项目(70871024)
关键词
非等间距
GM(1
1)模型
背景值
参数优化
non-equidistant
model GM(1
1)
background value
parameters optimization