-
Notifications
You must be signed in to change notification settings - Fork 46
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Make predict
easier?
#134
Comments
FWIW I'm now using this function:
which seems to do the job, but clearly is lacking any sort of uncertainty estimates / prediction intervals |
Yes, would be nice to define You'd need to modify this function
|
I'll have a look at this and try to get a PR together |
Okay I got stuck and am wondering whether I'm missing something:
Based on this, it seems to me that it's not in general possible to make a prediction from the model object and some new data - the only way to know which number in If this is correct, it seems to me there are two options here to allow predictions from new data:
I suppose (2) is potentially breaking to the extent that people have relied on the format of the Any comments appreciated - happy to implement either of these or both, or indeed none of them if I've missed something and there's a better way! |
Sorry for the late response — you're exactly right. I think Option2 is best. |
Fixed by #185 |
predict
currently raises an error, advising the user to estimate the model with thesave = :fe
kwarg and the access the fixed effects usingfe()
.If I'm not mistaken, this advise only covers the retrieval of fixed effects, but doesn't actually give a prediction. I'm currently doing:
which I believe produces the correct predictions from the model (do correct me if I'm getting something wrong here!)
Could this be turned into a
predict
method, which accepts thefem
object directly provided it has been estimated withsave = :fe
?The text was updated successfully, but these errors were encountered: