A new model selection criterion based on information geometry

被引：0

作者：

Liu, YH ^{[1
]}

Luo, SW ^{[1
]}

Li, AJ ^{[1
]}

Yu, HB ^{[1
]}

机构：

[1] Beijing Jiaotong Univ, Dept Comp Sci, Beijing 100044, Peoples R China

来源：

2004 7TH INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING PROCEEDINGS, VOLS 1-3 | 2004年

关键词：

model selection; information geometry; geometric complexity; geometric fitness;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

This paper presents a new model selection criterion based on information geometry-Information Geometric Model Selection Criterion (IGMSC) which is reparametrization invariant, and gives the proof IGMSC computes the geometric complexity of the model by regarding the model space as the manifold and estimates the model-data geometric fitness by using the divergence between the true distribution and the asymptotic distribution, enduing complexity and fitness with clear geometric significance. IGMSC gives the theoretic support of model selection in the framework of information geometry.

引用

页码：1562 / 1565

页数：4

共 50 条

[21] A strongly consistent information criterion for linear model selection based on M-estimation
Wu, Y
Zen, MM
PROBABILITY THEORY AND RELATED FIELDS, 1999, 113 (04) : 599 - 625
[22] Speaker model selection based on the Bayesian information criterion applied to unsupervised speaker indexing
Nishida, M
Kawahara, T
IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 2005, 13 (04): : 583 - 592
[23] Information geometry approach to the model selection of neural networks
Lv, Ziang
Luo, Siwei
Liu, Yunhui
Zheng, Yu
ICICIC 2006: FIRST INTERNATIONAL CONFERENCE ON INNOVATIVE COMPUTING, INFORMATION AND CONTROL, VOL 3, PROCEEDINGS, 2006, : 419 - +
[24] Marker selection by Akaike information criterion and Bayesian information criterion
Li, WT
Nyholt, DR
GENETIC EPIDEMIOLOGY, 2001, 21 : S272 - S277
[25] A new penalty-based criterion for model selection in regularized nonlinear models
Guerrero, E
Pizarro, J
Yáuzi, A
Galindo, P
COMPUTATIONAL METHODS IN NEURAL MODELING, PT 1, 2003, 2686 : 374 - 381
[26] An Akaike information criterion for model selection in the presence of incomplete data
Cavanaugh, JE
Shumway, RH
JOURNAL OF STATISTICAL PLANNING AND INFERENCE, 1998, 67 (01) : 45 - 65
[27] The weighted average information criterion for multivariate regression model selection
Wu, Tiee-Jian
Chen, Pinyuen
Yan, Yanjun
SIGNAL PROCESSING, 2013, 93 (01) : 49 - 55
[28] Information geometric model selection criterion and its application in cognition
Liu, Yun-Hui
Luo, Si-Wei
Lv, Zi-Ang
Huang, Hua
PROCEEDINGS OF 2006 INTERNATIONAL CONFERENCE ON MACHINE LEARNING AND CYBERNETICS, VOLS 1-7, 2006, : 2814 - +
[29] Bootstrap variants of the Akaike information criterion for mixed model selection
Shang, Junfeng
Cavanaugh, Joseph E.
COMPUTATIONAL STATISTICS & DATA ANALYSIS, 2008, 52 (04) : 2004 - 2021
[30] The reliability of the Akaike information criterion method in cosmological model selection
Tan, M. Y. J.
Biswas, Rahul
MONTHLY NOTICES OF THE ROYAL ASTRONOMICAL SOCIETY, 2012, 419 (04) : 3292 - 3303

← 1 2 3 4 5 →