On Fri, Apr 13, 2012 at 11:16 PM, Mathieu Blondel wrote:
>
> Ridge would be a candidate but there is a bit of work to be done to
> support warm start. Currently, Ridge supports two solvers: one based on
> scipy.linalg.solve in the dense case and another based on
> scipy.sparse.linalg.cg in the spa
On Fri, Apr 13, 2012 at 10:55 PM, Conrad Lee wrote:
> Thanks for the suggestion of how to change RFE.py to exploit warm_start.
> Should I add this feature to rfe.py and make a pull request? Or is this
> functionality too specialized?
>
+1 for a PR
Do you know of other fast estimators I can use
Le 13 avril 2012 15:55, Conrad Lee a écrit :
> Thanks for the suggestion of how to change RFE.py to exploit warm_start.
> Should I add this feature to rfe.py and make a pull request? Or is this
> functionality too specialized?
>
>> As we said in the issue you opened recently, SGDRegressor doesn't
2012/4/13 Conrad Lee :
> Thanks for the suggestion of how to change RFE.py to exploit warm_start.
> Should I add this feature to rfe.py and make a pull request? Or is this
> functionality too specialized?
>
>> As we said in the issue you opened recently, SGDRegressor doesn't monitor
>> convergence
Thanks for the suggestion of how to change RFE.py to exploit warm_start.
Should I add this feature to rfe.py and make a pull request? Or is this
functionality too specialized?
As we said in the issue you opened recently, SGDRegressor doesn't monitor
> convergence therefore warm_start won't make t
> Also, is the warm_start option relevant here? In theory, the coefficients
> from an estimator fit with N features could be used for fitting an
> estimator with N - 1 features. In practice though, RFE and RFECV might not
> copy the coef_ array between runs.
>
>
warm_start can be used but you hav
Hey, I've found that RFECV does a good job of selecting features, but that
it's quite slow. It is quicker if it's used with a fast estimator. For
example, I've found that SGDRegressor is much faster than SVR. I was
wondering any regressor is faster than SGDRegressor (for dense features).
Remembe
