Regularization for Autoregressive Processes and its Application to Long-memory Financial Time Series

被引:0
|
作者
Sun, Y. [1 ]
Lin, X. [1 ]
机构
[1] Univ Cincinnati, Dept Math Sci, Cincinnati, OH 45221 USA
关键词
NONCONCAVE PENALIZED LIKELIHOOD; ORACLE PROPERTIES; SELECTION; LASSO; ORDER;
D O I
暂无
中图分类号
C93 [管理学];
学科分类号
12 ; 1201 ; 1202 ; 120202 ;
摘要
This paper proposes a regularization framework for autoregressive processes based on penalized conditional likelihood. Simulation shows that the method has advantageous performances compared to the usual MLE when the coefficient profile is long and exhibits sparsity. A real financial data analysis is represented.
引用
收藏
页码:284 / 287
页数:4
相关论文
共 50 条