A Bayesian model-X knockoff procedure using parameter-expanded latent knockoffs from a Gaussian graphical model on covariates and modified spike-and-slab priors controls Bayesian FDR for variable selection in finite samples when the covariate distribution is known.
M., Kucukelbir, A., and McAuliffe, J
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
stat.ME 1years
2024 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Bayesian Controlled FDR Variable Selection via Parameter-Expanded Latent Knockoffs
A Bayesian model-X knockoff procedure using parameter-expanded latent knockoffs from a Gaussian graphical model on covariates and modified spike-and-slab priors controls Bayesian FDR for variable selection in finite samples when the covariate distribution is known.