Efficient MCMC algorithm for linear regression models that makes use of 'spike-and-slab' priors for some modest regularization on the amount of posterior probability for a subset of the coefficients.
Usage
engine_breg(
x,
iter = 10000,
nthread = getOption("ibis.nthread"),
type = "response",
...
)Arguments
- x
distribution()(i.e.BiodiversityDistribution) object.- iter
numericon the number of MCMC iterations to run (Default:10000).- nthread
numericon the number of CPU-threads to use for data augmentation.- type
The mode used for creating posterior predictions. Either making
"link"or"response"(Default:"response").- ...
Other none specified parameters passed on to the model.
Value
An Engine.
Details
This engine provides efficient Bayesian predictions through the
Boom R-package. However note that not all link and models functions are
supported and certain functionalities such as offsets are generally not
available. This engines allows the estimation of linear and non-linear
effects via the "only_linear" option specified in train.
References
Nguyen, K., Le, T., Nguyen, V., Nguyen, T., & Phung, D. (2016, November). Multiple kernel learning with data augmentation. In Asian Conference on Machine Learning (pp. 49-64). PMLR.
Steven L. Scott (2021). BoomSpikeSlab: MCMC for Spike and Slab Regression. R package version 1.2.4. https://CRAN.R-project.org/package=BoomSpikeSlab
See also
Other engine:
engine_bart(),
engine_gdb(),
engine_glm(),
engine_glmnet(),
engine_inla(),
engine_inlabru(),
engine_stan(),
engine_xgboost()
Examples
if (FALSE) {
# Add BREG as an engine
x <- distribution(background) |> engine_breg(iter = 1000)
}