【作者】 刘超； 贾知青；

【Author】 LIU Chao~1 JIA Zhi-qing~2 ( 1.Department of Applied Mathematics,School of Science,LIMB,Beihang University,Beijing 100083,China;2.Institute of Quantitative & Technical Economics,Chinese Academy of Social Scieneces 100732,China)

【机构】 北京航空航天大学应用数学系“数学、信息与行为”教育部重点实验室； 中国社科院数量经济与技术经济研究所；

【摘要】 针对居民消费行为中存在的地区差异和时间因子影响的研究,本文采用更一般形式的动态线性混合模型来进行建模,对其参数估计及其有关性质进行了讨论,采用 Monte Carlo 模拟说明了几种估计量的优良性和不同估计量之间的差异,最后我们采用此模型对我国居民消费行为进行了实证研究,并对所建议模型如解释能力和预测能力等进行了对比分析,最后提出改善居民消费结构的建议。更多还原

【Abstract】 On the research of region difference and time factor in residents consume behavior,we makes use of the more general form to the longitudinal data analysis,dynamic linear mixed models,to build this model,and discusses the characters of estimation of dynamic linear mixed models under the situations of the fixed effect models.By the simulation of Monte Carlo,we compare the effects(such as the validity)of series of parameter estimations we gain a- bove and have some conclusions.Finally we exploit the longer data on the residents’ consume and income based on this model,discuss the influence of the income diversities among urban citizens on the structures of consumption. Comparing with some different models,it explains the advantage of dynamic linear mixed models and gives some ad- vice.更多还原