/variance-priors

This repository contains the code to reproduce the results in Moran et al (2018).

Primary LanguageR

Variance prior forms for Bayesian variable selection

This repository contains R code to reproduce the results in Moran et al (2018).

This code uses the Spike-and-Slab Lasso which is implemented in the SSLASSO R package, available on CRAN.

Index

  • section_3-2.R

    This R script runs the simulation study in Section 3.2 and generates Figure 1.

  • section_6-6.R

    This R script runs the simulation study in Section 6.6 and generates Table 1 and Figure 2.

  • section_7.R

    This R script runs the analysis of the protein activity data of Clyde and Parmigiani (1998) which is detailed in Section 7.

  • scaledSSL

    This package implements the "scaled Spike-and-Slab Lasso" as described in Section 6.5 of Moran et al (2018). It is provided for the sole purpose of reproducing the results in section_6.6.R. If you would like to use the Spike-and-Slab Lasso with unknown variance, please use the package SSLASSO available on CRAN here. To install scaledSSL in order to run the code in "section_6.6.R", use the following commands:

library(devtools)
install_github("gemma-e-moran/variance-priors/scaledSSL")

References

  • Moran, G. E., Rockova, V. and George, E. I. (2018) "Variance prior forms for high-dimensional Bayesian variable selection'' [arXiv:1801.03019]
  • Clyde, M. A. and Parmigiani, G. (1998) "Protein construct storage: Bayesian variable selection and prediction with mixtures" Journal of Biopharmaceutical Statistics, 8, 431-443