摘要
针对近似非齐次指数的非等间距数据序列在建模过程中存在灰色微分方程和白化微分方程不匹配、建模后需要对白化微分方程解的时间响应式作累减还原问题,提出利用原始序列与其对应的1-IAGO构建灰色微分方程和白化微分方程相匹配的IANGM(1,1,k)模型新方法,该方法利用模型白化微分方程得到的时间响应函数是关于原始序列的表达式,无需累减还原就可以直接进行模拟预测。通过实例验证了新方法对递增和递减的近似非齐次指数的非等间距序列均具有非常好的模拟和短期预测效果,新方法进一步拓宽了灰色预测模型的适用范围。
Aiming at solve the problems for the non-equidistant data sequence of the non-homogeneous exponent,the grey differential equations and the whitening differential equations that exist in the modeling process do not match,and the modeling needs to reduce the time response of the solution of the whitening differential equations.A new method of IANGM(1,1,k)model using the original sequence and its corresponding 1-IAGO to construct grey differential equation and whitening differential equation.The time response function obtained by using the model whitening differential equation is about the original sequence.The expression can be directly simulated without needing of the inverse accumulating reduction process.Finally,an example is given to verify that the new method has very good simulation and short-term prediction effects for non-equidistant sequences with increasing and decreasing approximate non-homogeneous exponents.The new method further broadens the applicability of the grey prediction model.
作者
陈静
陈友军
CHEN Jing;CHEN You-jun(School of Mathematics and Information,China West Normal University,Nanchong 637002,China)
出处
《统计与信息论坛》
CSSCI
北大核心
2019年第12期16-21,共6页
Journal of Statistics and Information
基金
西华师范大学英才科研基金项目《网页编辑器代码优化算法研究》(17YC370)
西华师范大学教学改革与研究重点项目《基于互联网+的创新创业激励模式探索》(jgxmzd1824)
