A copula-based partition Markov procedure

被引:2
|
作者
Fernandez, M. [1 ]
Garcia, Jesus E. [2 ]
Gonzalez-Lopez, V. A. [2 ]
机构
[1] BM&FBOVESPA, Praca Antonio Prado, Sao Paulo, SP, Brazil
[2] Univ Estadual Campinas, Dept Stat, Rua Sergio Buarque de Holanda 651, BR-13083859 Campinas, SP, Brazil
基金
巴西圣保罗研究基金会;
关键词
Bayesian information criterion; copula distribution function; Markov chains; the curse of dimensionality; SELECTION;
D O I
10.1080/03610926.2017.1359291
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
The number of parameters needed to specify a discrete multivariate Markov chain grows exponentially with the order and dimension of the chain, and when the size of the database is not large enough, it is not possibly a consistent estimation. In this paper, we introduce a strategy to estimate a multivariate process with an order greater than the order achieved using standard procedures. The new strategy consists in obtaining a partition of the state space which is constructed from a combination of the partitions corresponding to the marginal processes and the partitions corresponding to the multivariate Markov chain.
引用
收藏
页码:3408 / 3417
页数:10
相关论文
共 50 条