Hi Paul (Bo?).
Which version of scikit-learn are you using and which version of numpy
and scipy? I guess scipy 0.16 numpy 1.10 and scikit-learn 0.16.2.
The errors for the newer versions of scipy and numpy are fixed in the
scikit-learn development version.
You shouldn't be concerned by these erro
Hi, here is what I got when I tired to install the sklearn.
Thanks
Paul
ERROR: sklearn.gaussian_process.tests.test_gaussian_process.test_2d
--
Traceback (most recent call last):
File
"/Users/paul/Library/Enthought/Canopy_64b
Hi, here is what I got when I tired to install the sklearn.
Thanks
Paul
ERROR: sklearn.gaussian_process.tests.test_gaussian_process.test_2d
--
Traceback (most recent call last):
File
"/Users/paul/Library/Enthought/Canopy_64b
You have sent this question multiple times to this mailing list. You might
consider responding to the other replies.
On Thu, Oct 29, 2015 at 12:28 PM, Shreyas Saligrama chandrakan <
ssali...@hawk.iit.edu> wrote:
> Hi,
>
> I would like to build a custom tree kernel with SVM's using scikit-learn.
>
Hi,
I would like to build a custom tree kernel with SVM's using scikit-learn.
How do i go about it to build this custom kernel. Any suggestions or
answers would be helpful.
Thanks.
--
_
scipy allows to perform the friedman test.
Orange has the tool to drawn the critical distance diagram.
And you can easily compute the critical distance using stats model:
from statsmodels.stats.libqsturng import qsturng
q_alpha = qsturng(1 - alpha, n_methods, np.inf) / np.sqrt(2)
cd = q_alpha * n
On 10/23/2015 05:17 PM, Ouwen Huang wrote:
>
> Hey Andy,
>
> I don't think it is so bad to code. Perhaps I could do a related
> projects repo first and then see if something similar can be moved in
> officially?
>
Yeah sounds good.
--
Sorry, don't know of a package. But it might be interesting for sklearn?
So that's a Nemenyi test?
https://en.wikipedia.org/wiki/Nemenyi_test
I never heard of that but it sounds interesting.
It seems a bit hard to interpret, though.
Also: does the diagram punt if the initial multiple comparison
(That is not to say you should rename it now. Just send a pull request
to multiclass and we discuss in the PR).
On 10/26/2015 10:04 AM, Al wrote:
Good afternoon,
The vanilla rakel classifier chain (with its dependencies) is (I
think) at last operational. However, I am not sure where I should
Hi Al.
They should probably go into "multiclass" at the moment. Which should
maybe be renamed "multilabel"?
Or "reduction" (which is really non-intuitive for non-experts)?
Cheers,
Andy
On 10/26/2015 10:04 AM, Al wrote:
Good afternoon,
The vanilla rakel classifier chain (with its dependenci
What do you mean by cost factor?
Have you implemented the tree kernel yourself?
On 10/26/2015 06:53 PM, Shreyas Saligrama chandrakan wrote:
Hi,
For my experiment, i would like to build a SVM tree kernel model by
tuning the weight of the contribution of trees and cost factor of the
trees. So,
Hi Mike.
This has been fixed in the development version (and the release
candidate which will be released imminently).
Best,
Andy
On 10/27/2015 05:59 PM, Michael Albert wrote:
Greetings!
When pickling a random forest fit, the storage requirements seem
disproportionately large.
It seems tha
Hi Shreyas.
Tree kernels are not implemented in scikit-learn, but it should be
possible to provide your own implementation as a callable to SVC.
I'm not sure how you want to use them in an unsupervised manner, though.
Do you want to do one-class SVM?
Andy
On 10/23/2015 10:27 PM, Shreyas Sali
Hi,
Do you guys know any tool to generate CDdiagram - in order to evaluate the
difference of performance of sklearn classifiers?
http://theoval.cmp.uea.ac.uk/matlab/critdiff/cd1.png​
There is a R package called performanceEstimation which has
a CDdiagram implementation, but it uses an specific
14 matches
Mail list logo
