PASCAL - Pattern Analysis, Statistical Modelling and Computational Learning

On the consistency of sequentially normalized least squares
Daniel Schmidt and Teemu Roos
In: The Third Workshop on Information Theoretic Methods in Science and Engineering (WITMSE), 18-20 August 2010, Tampere, Finland.

Abstract

We examine the Sequentially Normalized Least Squares (SNLS) criterion for linear regression model selection. In particular, we present (i) a simplified formula for computing the SNLS score, (ii) an asymptotic representation of the SNLS score even in the case of model misspecification, and (iii) a proof of the consistency of SNLS as a model selection tool.

PDF - Requires Adobe Acrobat Reader or other PDF viewer.
EPrint Type:Conference or Workshop Item (Invited Talk)
Project Keyword:Project Keyword UNSPECIFIED
Subjects:Computational, Information-Theoretic Learning with Statistics
ID Code:7285
Deposited By:Teemu Roos
Deposited On:16 March 2011