摘要
灰色 GM( 1 ,1 )模型对高增长指数序列拟合常常产生滞后误差 ,作者认为 GM( 1 ,1 )模型中背景值构造方法是影响其精度和适应性的关键因素 .从此角度出发 ,对背景值构造方法进行研究 ,重构了一个表达形式简洁、计算简单、适应性极强的背景值计算公式 .新的背景值计算公式的一个显著特点是它使 GM( 1 ,1 )模型具有对建模结果进行优化的能力 ,能获得最佳的拟合和预测精度 .它使 GM( 1 ,1 )模型同时适应于低增长指数序列和高增长指数序列建模 ,它是提高 GM( 1 ,1 )模型精度和适应性的关键技术 .算例结果的精度充分说明了它的有效性 .
In this paper, for the problem that grey GM(1,1) model often makes delay error for fitting high growth index series, we point out that the structure method of background value in grey GM(1,1) model has an important influence on the precision and adaptability of the model. From this point of view, the structure method of background value is researched, a simple calculating formula of background value is reestablished which has strong adaptability. An obvious feature of the new calculating formula of background value is that it makes GM(1,1) model have the ability to optimize the modeling results, the best fitting and forecasting precision can be obtained with it. The new formula is suited to building GM(1,1) model for both low growth index series and high growth index series. It is the key technology of enhancing GM(1,1) model precision and adaptability. The calculating precision of examples show this new formula is very effective.
出处
《系统工程理论与实践》
EI
CSCD
北大核心
2000年第4期98-103,共6页
Systems Engineering-Theory & Practice
