Package: DPw
Type: Package
Title: Semi- and nonparametric Bayesian Subset Selection Procedures
Version: 1.1.3
Date: 2014-12-09
Author: Yumi Kondo
Maintainer: Yumi Kondo <y.kondo@stat.ubc.ca>
Description: Implement the Bayesian subset selection procedure proposed in the paper titled "Bayesian Subset Selection Procedures with an Application to Lumber Strength Properties" by Yumi Kondo, James V. Zidek, Carolyn Taylor, and Constance van Eeden. This procedure models the lumber strengths with Dirichlet processes (DP) with the Weibull components and finds, with pre-specified certainty, the smallest subset that contains those species having the smallest alpha-th quantile of strength (i.e., the weakest species). We consider two models for the strengths: [1] A semi-parametric model which assumes that strengths are from the DP mixture of Weibulls; [2] the nonparametric model which assumes that probability measures of strengths are from DP with Weibull base distribution. Either model can be used to estimate the posterior probability that a subset contains the weakest species.
License: GPL (>= 2)
Depends: fOptions
Packaged: 2014-12-11 20:09:24 UTC; yumik091186
NeedsCompilation: yes
Repository: CRAN
Date/Publication: 2014-12-12 10:56:30
Built: R 3.1.2; x86_64-apple-darwin10.8.0; 2014-12-13 13:14:23 UTC; unix
Archs: DPw.so.dSYM
