search

merliseclyde / BAS

BAS R package for Bayesian Model Averaging and Variable Selection
https://merliseclyde.github.io/BAS/
GNU General Public License v3.0
42 stars 16 forks source link

bestmodel documentation and use #42

Open A-gata opened 5 years ago

A-gata commented 5 years ago

A clear and concise description of what you would like to see added to BAS. I think the bestmodel argument is not clearly explained if one wants to modify it, would it be possible to include an example of using it with something other than the default?

I understand it as being able to specify which model should come first but it's not clear how and I haven't been able to determine that.

merliseclyde commented 1 year ago

Thank you! I had overlooked this issue thinking it was for adding more documentation for extracting the best model (which depends on the loss function - but now I see that this is really for added documentation for the input argument bestmodel I believe.

That should probably be renamed as initial_model as it is just the model used to start sampling, and is a way to re-start sampling at say the last found highest probability model or other "best" models. It can also be use to fit just a single model, as is the case for extracting the Median Probability Model.

ToDo: Add more documentation in the help files and vignette