| boral | R Documentation |
This list below is written in terms of what is deemed to be most important to least important changes =P
A new argument lasso.lambda.scale has been included as part of control in the main rpql function. This scales the penalty on the group random effects, when the lasso or adaptive lasso penalty is used, by 1/sqrt(no. of clusters) so that it is on the same scale as the fixed effects penalty; Thanks to Sarbesh Pandeya for this suggestion. Defaults to TRUE.
Modified some of the examples in response to the above change.
Fixed a bug issue when trial.size > 1.