A prequential approach to regression estimation

Author

Modha, Dharmendra S. ; Masry, Elias

Author_Institution

IBM Almaden Res. Center, San Jose, CA, USA

fYear

1997

fDate

29 Jun-4 Jul 1997

Firstpage

404

Abstract

Prequential model selection is a data-driven methodology for selecting between rival models on the basis of their predictive ability where the predictive ability of a model is measured by its accumulated prediction error on a given set of observations. Given i.i.d. observations, we propose a regression estimator-based on neural networks-that selects the number of “hidden units” using prequential model selection, and establish a rate of convergence for the statistical risk of the proposed estimator

Keywords

convergence of numerical methods; data structures; estimation theory; neural nets; parameter estimation; prediction theory; statistical analysis; accumulated prediction error; convergence rate; data-driven methodology; hidden units; i.i.d. observations; neural networks; parametric model; predictive ability; prequential model selection; regression estimation; sequence; statistical risk; Approximation error; Convergence; Data engineering; Electric variables measurement; Estimation error; Home computing; Neural networks; Parametric statistics; Predictive models; Random variables;

fLanguage

English

Publisher

ieee

Conference_Titel

Information Theory. 1997. Proceedings., 1997 IEEE International Symposium on

Conference_Location

Ulm

Print_ISBN

0-7803-3956-8

Type

conf

DOI

10.1109/ISIT.1997.613341

Filename

613341