Forum: developers
Monitor Forum | Start New ThreadRE: "terms" and model.frame [ Reply ] By: Ott Toomet on 2011-12-25 06:38 | [forum:5363] |
Good. Should look what the other multi-equation models are doing. |
RE: "terms" and model.frame [ Reply ] By: Arne Henningsen on 2011-12-23 07:57 | [forum:5357] |
Your idea with model.frame.selection( object, part , ... ) sounds like a good idea and is consistent with some other methods for "selection" objects. I have no better idea. |
"terms" and model.frame [ Reply ] By: Ott Toomet on 2011-12-06 08:27 | [forum:5264] |
Currently model.frame.selection returns terms for selection equation (as attribute). This seems to work, but I feel it a little unclean approach. In this way it is not possible to recover all the formulas from the model.frame, and it does not allow the possibility that different equations are evaluated in different environments. None of these are necessary, and currently seems like none of the standard methods care too much about it. But you never know.. Are there a cleaner approach? What do the other multi-equation models do? A way to go were to specify the model frame as: model.frame.selection <- function(object, part=c("full", "selection", "outcome1", ...). Ideas? |