Bayesian wavelet networks for nonparametric regression.
Radial wavelet networks have recently been proposed as a method for nonparametric regression. In this paper we analyze their performance within a Bayesian framework. We derive probability distributions over both the dimension of the networks and the network coefficients by placing a prior on the deg...
| Main Authors: | , |
|---|---|
| Format: | Journal article |
| Language: | English |
| Published: |
IEEE
2000
|
| _version_ | 1797063162597474304 |
|---|---|
| author | Holmes, C Mallick, B |
| author_facet | Holmes, C Mallick, B |
| author_sort | Holmes, C |
| collection | OXFORD |
| description | Radial wavelet networks have recently been proposed as a method for nonparametric regression. In this paper we analyze their performance within a Bayesian framework. We derive probability distributions over both the dimension of the networks and the network coefficients by placing a prior on the degrees of freedom of the model. This process bypasses the need to test or select a finite number of networks during the modeling process. Predictions are formed by mixing over many models of varying dimension and parameterization.We show that the complexity of the models adapts to the complexity of the data and produces good results on a number of benchmark test series. |
| first_indexed | 2024-03-06T20:55:53Z |
| format | Journal article |
| id | oxford-uuid:393c7bfc-5e5f-4b1d-a933-6cdb13fd4ce3 |
| institution | University of Oxford |
| language | English |
| last_indexed | 2024-03-06T20:55:53Z |
| publishDate | 2000 |
| publisher | IEEE |
| record_format | dspace |
| spelling | oxford-uuid:393c7bfc-5e5f-4b1d-a933-6cdb13fd4ce32022-03-26T13:54:25ZBayesian wavelet networks for nonparametric regression.Journal articlehttp://purl.org/coar/resource_type/c_dcae04bcuuid:393c7bfc-5e5f-4b1d-a933-6cdb13fd4ce3EnglishSymplectic Elements at OxfordIEEE2000Holmes, CMallick, BRadial wavelet networks have recently been proposed as a method for nonparametric regression. In this paper we analyze their performance within a Bayesian framework. We derive probability distributions over both the dimension of the networks and the network coefficients by placing a prior on the degrees of freedom of the model. This process bypasses the need to test or select a finite number of networks during the modeling process. Predictions are formed by mixing over many models of varying dimension and parameterization.We show that the complexity of the models adapts to the complexity of the data and produces good results on a number of benchmark test series. |
| spellingShingle | Holmes, C Mallick, B Bayesian wavelet networks for nonparametric regression. |
| title | Bayesian wavelet networks for nonparametric regression. |
| title_full | Bayesian wavelet networks for nonparametric regression. |
| title_fullStr | Bayesian wavelet networks for nonparametric regression. |
| title_full_unstemmed | Bayesian wavelet networks for nonparametric regression. |
| title_short | Bayesian wavelet networks for nonparametric regression. |
| title_sort | bayesian wavelet networks for nonparametric regression |
| work_keys_str_mv | AT holmesc bayesianwaveletnetworksfornonparametricregression AT mallickb bayesianwaveletnetworksfornonparametricregression |