vbglmss: Variational approximation of Spike and Slab GLM

In antiphon/vbss: Spike and Slab generalized linear regression (GLM) using variational approximation

Description

The function fits the GLM model in Bayesian framework with some covariate coefficients having the spike and slab prior.

Usage

  vbglmss(formula, family = gaussian, data = NULL,
    intercept = FALSE, y = NULL, X = NULL, Z = NULL, ...)

Arguments

formula	an object of class formula. See Details.
family	the likelihood family to be used in the model, see family.
data	an optional data frame, list or environment (or object coercible by as.data.frame to a data frame) containing the variables in the model. If not found in data, the variables are taken from environment(formula), much like in glm.
intercept	should the intercept term be explicitely added.
y	Alternative for formula&data: The response
X	Alternative for formula&data: The SS covariates
Z	Alternative for formula&data: The non-SS (regular Gaussian) covariates
method	Use 'simultaneous' or 'sequential'. See Details.
verbose	print additional runtime messages
eps	the approximation loop convergence threshold

Details

The formula should use the special tags 'ss' for those covariates that are set the spike and slab prior. Without it, we assume regular Gaussian prior.

Using the y,X,Z variables is useful to avoid overhead in formula parsing when data is large. This is then equal to the formula y ~ ss(X) + Z

Value

An object of class vbglmss.

antiphon/vbss documentation built on May 10, 2019, 12:22 p.m.