摘要
针对传统MGM(1,m)模型未能反映多变量系统时滞效应的问题,文章根据灰色系统建模思想构建了多变量时滞离散MGM(1,m,τ)模型,研究该模型的建模机理、建模过程,并给出时滞参数的计算方法。最后,通过对中国城镇居民人均可支配收入和消费支出的建模预测实例,探讨其滞后效应,验证该模型的有效性。结果发现时滞离散MGM(1,m,τ)模型具有较高的建模精度,能够较好地模拟预测含时滞特征的小样本多变量数据。
Aiming at the fact that the traditional MGM(1, m) model fails to reflect the time-delay effect of multivariable systems, this paper constructs a multivariable discrete MGM(1,m,τ) model based on the grey system modeling theory, makes a study on the modeling mechanism and process of the model, and also gives the calculation method of time-delay parameters. Finally, by modeling and forecasting the per capita disposable income and consumption expenditure of urban residents in China, the paper discusses the hysteresis effect and verifies the validity of the model. The results show that MGM(1, m, participle) model has comparatively higher modeling accuracy, able to well simulate and predict small sample multivariate data with time-delay characteristics.
作者
熊萍萍
张悦
邢卓
陈峰
Xiong Pingping;Zhang Yue;Xing Zhuo;Chen Feng(College of Mathematics and Statistics,Nanjing University of Information Science and Technology, Nanjing 210044, China;China Institute of Manufacturing Development, Nanjing University of Information Science and Technology, Nanjing 210044, China)
出处
《统计与决策》
CSSCI
北大核心
2019年第8期18-22,共5页
Statistics & Decision
基金
国家自然科学基金资助项目(71701105
41505118)
国家社会科学基金重大项目(17ZDA092)
教育部人文社会科学研究青年基金项目(17YJC630182
17YJC630123)
江苏省高校哲学社会科学研究重点项目(2018SJZDI111)
中国制造业发展研究院开放课题(SK20140090-13)
大学生实践创新训练计划国家级项目(201710300041)