A Bayesian approach to flexible modeling of multivariable response functions
This paper presents a Bayesian approach to empirical regression modeling in which the response function is represented by a power series expansion in Hermite polynomials. The common belief that terms of low degree will reasonably approximate the response function is reflected by assigning prior distributions that exponentially downweight the coefficients of high-degree terms. The model thus includes the complete series expansion. A useful property of the Hermite expansion is that it can be easily extended to handle models with several explanatory variables.
Year of publication: |
1990
|
---|---|
Authors: | Steinberg, David M. |
Published in: |
Journal of Multivariate Analysis. - Elsevier, ISSN 0047-259X. - Vol. 34.1990, 2, p. 157-172
|
Publisher: |
Elsevier |
Keywords: | Bayesian linear model empirical modeling Hermite polynomials polynomial regression smoothing |
Saved in:
Online Resource
Saved in favorites
Similar items by person
-
Thoughts on Data Science in business and industry
Steinberg, David M., (2020)
-
Item response function in antagonistic situations
Turetsky, Vladimir, (2020)
-
Computer experiments : a review
Levy, Sigal, (2010)
- More ...