This would require rolling our own implementation and this would be a non-trivial amount of work. Preliminary research also shows that Bayesian feature selection is very intensive and of most benefit to domains where there are large numbers of features to choose from (e.g. genomics).
This would require rolling our own implementation and this would be a non-trivial amount of work. Preliminary research also shows that Bayesian feature selection is very intensive and of most benefit to domains where there are large numbers of features to choose from (e.g. genomics).
Putting on hold for now.