The issue#39 bug was about incorrectly handling very few or no models. It was a very revealing bug and as a chain reaction I had to 1) fix both CV routines, 2) change both plot() and predict() for not always finding .1se models, 3) refactor (rather massively) changes to DMRnet, glamer and SOSnet
The issue#39 bug was about incorrectly handling very few or no models. It was a very revealing bug and as a chain reaction I had to 1) fix both CV routines, 2) change both plot() and predict() for not always finding .1se models, 3) refactor (rather massively) changes to DMRnet, glamer and SOSnet