Some notes on Rissanen's stochastic complexity

被引：22

作者：

Qian, GQ ^{[1
]}

Kunsch, HR

机构：

[1] La Trobe Univ, Sch Stat, Bundoora, Vic 3083, Australia

[2] ETH Zentrum SOL F4, Seminar Stat, CH-8092 Zurich, Switzerland

来源：

IEEE TRANSACTIONS ON INFORMATION THEORY | 1998年 / 44卷 / 02期

关键词：

Fisher information; model selection; robust regression; stochastic complexity; two-part codes;

D O I：

10.1109/18.661521

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

A new version of stochastic complexity for a parametric statistical model is derived, based on a class of two-part-codes. We show that choosing the quantization in the first step according to the Fisher information is optimal and we compare our approach to a recent result of Rissanen, An application to robust regression model selection is presented.

引用

页码：782 / 786

页数：5

共 11 条

[1]

Bentley J. L., 1976, Information Processing Letters, V5, P82, DOI 10.1016/0020-0190(76)90071-5

[2]

ELIAS P, 1975, IEEE T INFORM THEORY, V21, P194, DOI 10.1109/TIT.1975.1055349

[3]

HALL P, 1988, BIOMETRIKA, V75, P705

[4]

Hampel F. R., 1986, ROBUST STAT APPROACH