平滑转换向量乘积误差模型及对中国股市高频交易数据的应用  

Smooth Transition Vector Multiplicative Error Model:An Application to High Frequency Data in Chinese Stock Market

作　　者：李俊功 鲁万波[3] LI Jun-gong;LU Wan-bo(School of Business,Jiangsu University of Technology,Jiangsu Changzhou 213001,China;School of Economics and Finance,Tibet University,Tibet Lhasa 85000,China;School of Statistics,Southwest University of Finance and Economics,Sichun Chengdu 611130,China)

机构地区：[1]江苏理工学院商学院,江苏常州213001 [2]西藏大学财经学院,西藏拉萨850000 [3]西南财经大学统计学院,四川成都611130

出　　处：《数理统计与管理》2018年第6期1125-1140,共16页Journal of Applied Statistics and Management

基　　金：西藏大学青年科研培育基金(ZDPJSK1718);国家自然科学基金面上项目(71771187),国家自然科学基金青年项目(71101118);教育部新世纪优秀人才支持计划项目(NCET-13-0961)的资助

摘　　要：基于平滑转换理论本文提出了平滑转换向量乘积误差模型（STVMEM）。给出了模型极大似然估计迭代算法的初值选取方法和参数似然估计相结合的参数估计具体过程。参数估计具有一致性和渐进正态性。模拟结果表明参数估计具有良好的有限样本性质。利用信息分解形式的STVMEM对中国股市成交量、持续期和波动率之间关系进行分析,发现成交量冲击与波动率、波动冲击与持续期之间具有非线性关系。This paper proposes a Smooth Transition Vector Multiplicative Error Model （STVMEM）,which is based on regime transition theory. The specific process of parameter estimation combined with initial value selection is given in maximum likelihood estimation iterative algorithm for STVMEM. Sim- ulation results demonstrate the well behavior of our proposed method in finite samples. Based on the proposed STVMEM, in the form of information decomposition, we examine the relations among volume, duration and volatility. Results shows that there is nonlinear relationship between volume shocks and volatility, and just as the relationship between volatility shocks and duration.

关 键 词：平滑转换向量乘积误差模型 极大似然估计 非线性 高频数据 

分 类 号：F224.0[经济管理—国民经济] O212[理学—概率论与数理统计]