Re: [Scikit-learn-general] "MemoryError() in 'sklearn.tree._tree.Tree._resize' ignored"

Thanks! What should be the proper behavior when I run the script I wrote? I tried uninstalling and reinstalling from pip, but no success (same error), so I assume I have to build to get the bleeding edge update. Probably serves me right for using Windows, but I've been having trouble building sc

Re: [Scikit-learn-general] Sparse Gradient Boosting & Fully Corrective Gradient Boosting

Key advantage of using RuleFit [1] -- striking that they didnt cite it btw -- is that if you add the original features your model can a) better incorporate additive effects and b) extrapolate, a limitation of any tree-based method like GBRT or RF. [1] http://statweb.stanford.edu/~jhf/R-RuleFit.htm

Re: [Scikit-learn-general] Sparse Gradient Boosting & Fully Corrective Gradient Boosting

2014-09-21 10:46 GMT+02:00 Mathieu Blondel : > > > On Sun, Sep 21, 2014 at 1:55 AM, Olivier Grisel > wrote: >> >> On a related note, here is an implementeation of Logistic Regression >> applied to one-hot features obtained from leaf membership info of a >> GBRT model: >> >> >> http://nbviewer.ipyt

Re: [Scikit-learn-general] Issues for 1.0

On 09/22/2014 12:46 PM, Arnaud Joly wrote: Would it be possible that the issue with labels is https://github.com/scikit-learn/scikit-learn/issues/2451 ? That is certainly another complication, but I don't think that was the main issue. We definitely need to fix that one, too. Best regards,

Re: [Scikit-learn-general] Issues for 1.0

Would it be possible that the issue with labels is https://github.com/scikit-learn/scikit-learn/issues/2451 ? Best regards, Arnaud On 21 Sep 2014, at 19:12, Andy wrote: > Hi all. > I remember that we had a couple of things we wanted to do for 1.0, but I > didn't really find a good list. > I k

Re: [Scikit-learn-general] Backward compat policy in utils

On 09/22/2014 11:36 AM, Lars Buitinck wrote: > 2014-09-22 11:32 GMT+02:00 Andy : >> PyStruct uses >> minimum_spanning_tree > I removed that a few weeks ago, because nothing in scikit-learn was > using it. You can get it from scipy.sparse.csgraph. Thanks. I think when I wrote that it was pretty fres

Re: [Scikit-learn-general] Backward compat policy in utils

2014-09-22 11:32 GMT+02:00 Andy : > PyStruct uses > minimum_spanning_tree I removed that a few weeks ago, because nothing in scikit-learn was using it. You can get it from scipy.sparse.csgraph. -- Meet PCI DSS 3.0 Complia

Re: [Scikit-learn-general] "MemoryError() in 'sklearn.tree._tree.Tree._resize' ignored"

2014-09-22 11:30 GMT+02:00 Lars Buitinck : > 2014-09-21 6:42 GMT+02:00 c TAKES : >> looks like it is version 0.15.2 > > Confirmed. https://github.com/scikit-learn/scikit-learn/issues/3684 -- Meet PCI DSS 3.0 Compliance Re

Re: [Scikit-learn-general] Backward compat policy in utils

PyStruct uses minimum_spanning_tree check_random_state gen_even_slices shuffle check_arrays (this one is just for a backport of train_test_split, which I probably don't need any more) Cheers, Andy On 09/15/2014 03:40 PM, Mathieu Blondel wrote: lightning is using the following utils: - check

Re: [Scikit-learn-general] "MemoryError() in 'sklearn.tree._tree.Tree._resize' ignored"

2014-09-21 6:42 GMT+02:00 c TAKES : > looks like it is version 0.15.2 Confirmed. > I figured it must have been something I did wrong rather than a bug because > I think this is the same version as the last time I used sklearn decision > trees (random forest) and don't remember having a problem.

Re: [Scikit-learn-general] Backward compat policy in utils

2014-09-15 15:40 GMT+02:00 Mathieu Blondel : > lightning is using the following utils: > > - check_random_state > - safe_sparse_dot > - shuffle > - safe_mask > - sklearn.utils.testing.* seqlearn is using * atleast2d_or_csr * check_random_state * logsumexp * safe_sparse_dot so it's broken no