Presentation 2003/7/21
Statistical Inference in Singular Models
Kenji FUKUMIZU, Satoshi KURIKI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Parametric statistical models are often used for inference and learning from data. A statiscal model is called "singlar" if the model, which is regarded as a subset of a functional space, has a non-smooth point. There are many singular models among popular statistical models such as mixture models, ARMA, and HMM. If a model has such singularity, we see many interesting behavior on the learning machine or statistical model. After reviewing the standard theory on the models without singularities, this paper discusses known theoretical results on statistical behavior of the parameter obtained by estimation or learning.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Statistical inference / singular model / likelihood ratio / tangent cone
Paper # NC2003-27
Date of Issue

Conference Information
Committee NC
Conference Date 2003/7/21(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Statistical Inference in Singular Models
Sub Title (in English)
Keyword(1) Statistical inference
Keyword(2) singular model
Keyword(3) likelihood ratio
Keyword(4) tangent cone
1st Author's Name Kenji FUKUMIZU
1st Author's Affiliation Institute of Statistical Mathematics()
2nd Author's Name Satoshi KURIKI
2nd Author's Affiliation Institute of Statistical Mathematics
Date 2003/7/21
Paper # NC2003-27
Volume (vol) vol.103
Number (no) 227
Page pp.pp.-
#Pages 6
Date of Issue