基于阈值惩罚自回归模型的消费者信心指数的统计研究  

Statistical research on consumer confidence index based on threshold penalty autoregressive model

作　　者：袁晓惠 王慧雪 向雪飞 陈尧 YUAN Xiao-hui;WANG Hui-xue;XIANG Xue-fei;CHEN Yao(School of Mathematics and Statistics,Changchun University of Technology,Changchun 130012,China)

机构地区：[1]长春工业大学数学与统计学院,吉林长春130012

出　　处：《吉林师范大学学报（自然科学版）》2020年第3期44-50,共7页Journal of Jilin Normal University:Natural Science Edition

基　　金：国家自然科学基金项目(11571051,11671054)。

摘　　要：介绍了阈值自回归模型,通过加入惩罚达到自回归定阶,并给出了相应算法,不仅能选出自回归参数,而且能估计阈值点.模拟研究发现,带有惩罚的阈值自回归模型的估计效果更好.实证选取了我国2011—2018年消费者信心指数的经济数据,通过阈值自回归模型和阈值惩罚自回归模型计算得到消费者信心指数的阈值点均为106,即消费者信心指数在106处存在变点,说明消费者趋于乐观状态.It was introduced a threshold autoregressive model,which achieved autoregressive ordering by adding a penalty function,and gave the corresponding algorithm,which can not only select autoregressive parameters,but also estimate threshold points.Simulation studies have found that the threshold autoregressive model with penalty was better.The empirical selection of economic data of China's consumer confidence index from 2011 to 2018,the threshold of the consumer confidence index calculated by the threshold autoregressive model and the threshold penalty autoregressive model was 106,that was,the consumer satisfaction index changed at 106 points,indicating that consumers tend to be optimistic.

关 键 词：阈值惩罚自回归模型 消费者信心指数 阈值点 相依误差 

分 类 号：O212[理学—概率论与数理统计]