Efficient algorithm for solving ultra-sparse regularized regression models using a variational Bayes algorithm with a spike (l0) prior. Algorithm is solved on a path, with coordinate updates, and is capable of generating very sparse models. There are very general model diagnostics for controling type-1 error included in this package.
|Date of publication||2014-06-05 22:50:33|
|Maintainer||Benjamin Logsdon <firstname.lastname@example.org>|