Generalized weighted additive models based on distribution functions

초록

In this paper, a new form of generalized additive models is proposed. The proposed models consist of a distribution function of the mean response and a weighted linear combination of distribution functions of covariates. This form does not make any structural problems on linking the mean response and covariates. Markov chain Monte Carlo methods are used to estimate the parameters within a Bayesian framework. (c) 2007 Elsevier B.V. All rights reserved.

키워드