摘要
针对经典NGM(1,1,k)在背景值的影响下模型精度(拟合精度与预测精度)不高这一现状,结合复化求积公式中的复化梯形公式,推导了一种新的背景值优化公式.通过7类测试数据和2类实际数据的验证表明:推导的NGM(1,1,k)背景值优化公式显著地提高了NGM(1,1,k)的模型精度和实用性.
In view of the fact that the model accuracy(fitting accuracy and prediction accuracy) of classical NGM(1,1,k) is not so high under the background value, in this paper,a new background value optimization formula is derived based on the composite trapezoid formula. The validation of 7 kinds of test data and 2 kinds of actual data shows that the formula deduced in this paper significantly improves the accuracy and practicability of NGM(1,1,k) model.
作者
陈七榕
罗飞
何梦秋
CHEN Qi-rong;LOU Fei;HE Meng-qiu(School of Software Engineering,Chengdu University of Information Technology,Chengdu 610225,China)
出处
《数学的实践与认识》
北大核心
2019年第2期225-233,共9页
Mathematics in Practice and Theory
关键词
NGM(1
1
k)
背景值优化
复化梯形公式
NGM(1,1,k)
optimization of Background value
composite trapezoid formula