Misparametrization subsets for penalized least squares model selection

Identifying a model by the penalized contrast procedure, we give an analytical estimation of misfitting subsets in the specific case of a least squares contrast. Then, specifying the statistical model, this allows to determine penalization rates ensuring a consistent identification. Applications are...

Full description

Saved in:
Bibliographic Details
Published inStatistical inference for stochastic processes : an international journal devoted to time series analysis and the statistics of continuous time processes and dynamic systems Vol. 17; no. 3; pp. 283 - 294
Main Authors Guyon, Xavier, Hardouin, Cécile
Format Journal Article
LanguageEnglish
Published Dordrecht Springer Netherlands 01.10.2014
Subjects
Mathematics
Mathematics and Statistics
Probability Theory and Stochastic Processes
Statistical Theory and Methods
Model selection
Misfitting subsets
Penalized least squares contrast
Online AccessGet full text

Cover

More Information
Summary:Identifying a model by the penalized contrast procedure, we give an analytical estimation of misfitting subsets in the specific case of a least squares contrast. Then, specifying the statistical model, this allows to determine penalization rates ensuring a consistent identification. Applications are given to time series and geostatistical identification.
ISSN:1387-0874
1572-9311
DOI:10.1007/s11203-014-9100-y