On Wed, Jul 23, 2014 at 4:47 AM, Peter Prettenhofer <
peter.prettenho...@gmail.com> wrote:
>
> An alternative is to use a GradientBoostingRegressor with quantile loss to
> generate prediction intervals (see [1]) -- only for the keen - i've once
> used that unsuccessfully in a Kaggle comp. Its not a confidence score
> though -- it can only tell you if its within a band.
>
Indeed, the notion of quantile regression seems to differ from confidence
intervals. After all, we could also ask for a confidence interval on the
quantile predictions.
Besides my PR as mentioned by Peter, there is also this issue for RF and
bagging estimators
https://github.com/scikit-learn/scikit-learn/issues/3271
For these estimators, it is trivial to compute the empirical variances.
# I was kind of hoping someone would implement it during the spring ;-)
It will be nice to have have prediction variances in several estimators and
with a consistent API.
Mathieu
------------------------------------------------------------------------------
Want fast and easy access to all the code in your enterprise? Index and
search up to 200,000 lines of code with a free copy of Black Duck
Code Sight - the same software that powers the world's largest code
search on Ohloh, the Black Duck Open Hub! Try it now.
http://p.sf.net/sfu/bds
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
