面板数据中均值多变点和方差多变点的估计  

Multiple Change-point Estimation in the Mean and Variance of Panel Data Model

作　　者：徐小平[1] 刘君 杨倩男 李拂晓 XU Xiaoping;LIU Jun;YANG Qiannan;LI Fuxiao(School of Science,Xi'an University of Technology,Xi'an 710054)

机构地区：[1]西安理工大学理学院,西安710054

出　　处：《工程数学学报》2023年第2期251-264,共14页Chinese Journal of Engineering Mathematics

基　　金：国家自然科学基金(11801438);陕西省自然科学基础研究计划(2023-JC-YB-058);陕西省创新能力支撑计划(2020PT-023).

摘　　要：面板数据中的变点问题是金融和计量经济学中的一类重要课题。现将局部形态识别的二元分割算法推广到面板数据上,研究了面板数据中均值多变点和方差多变点的估计问题,同时得到了变点个数及变点位置的估计,证明了估计量的相合性。Monte Carlo模拟表明基于局部形态识别的二元分割算法对这两类变点的估计效果均优于二元分割算法。最后将提出的算法分别应用于不同的股市周指数数据,说明了算法的有效性。Change-point problem in panel data is an important topic in finance and econometric.A multiple change-point estimation procedure based on the shape-based BS algorithm is extended to panel data,multiple change-point estimation in the mean and variance of panel data model is studied.The number and location of the change-points are also obtained,and the consistency of the change point estimators is proved.Monte Carlo simulation shows that the shape-based BS algorithm outperforms the BS algorithm in estimating both types of change points.Finally,this method is applied to different weekly index data of the stock market,which shows the effectiveness of the algorithm.

关 键 词：面板数据 多变点估计 二元分割算法 局部形态识别 

分 类 号：O212.1[理学—概率论与数理统计]