Alex, I am afraid some codes have been broken...
In [127]: X = [[0], [1]]
In [129]: Y = [0, 1]
In [130]: clf.fit(X, Y)
Out[130]:
SVC(C=1.0, cache_size=200, coef0=0.0, degree=3, gamma=0.0, kernel='linear',
probability=False, scale_C=False, shrinking=True, tol=0.001)
In [132]: clf.decision_function(X)
Out[132]:
array([[-0.5],
[ 0.5]])
In [133]: X2 = [[1], [0]]
In [134]: Y2 = [1, 0]
In [135]: clf.fit(X2, Y2)
Out[135]:
SVC(C=1.0, cache_size=200, coef0=0.0, degree=3, gamma=0.0, kernel='linear',
probability=False, scale_C=False, shrinking=True, tol=0.001)
In [137]: clf.decision_function(X2)
Out[137]:
array([[ 0.5],
[-0.5]])
On Mon, Mar 26, 2012 at 8:59 PM, Alexandre Gramfort <
[email protected]> wrote:
> hi,
>
> from what I remember we fixed the random label ordering problem
> at least for the 2 classes case.
>
> can you check that things behave fine and the same way with Y = [0,
> 1] and Y = [1, 0]?
>
> Alex
>
> On Mon, Mar 26, 2012 at 5:00 AM, xinfan meng <[email protected]> wrote:
> > I use the following codes to obtain decision values for SVC classifier
> clf.
> >
> >
> -----------------------------------------------------------------------------------------------
> >
> > In [5]: >>> clf = svm.SVC()
> >
> > In [23]: >>> X = [[0], [1], [2]]
> >
> > In [24]: >>> Y = [0, 1, 2]
> >
> > In [25]: clf.fit(X, Y)
> > Out[25]:
> > SVC(C=1.0, cache_size=200, coef0=0.0, degree=3, gamma=1.0, kernel='rbf',
> > probability=False, scale_C=False, shrinking=True, tol=0.001)
> >
> > In [26]: clf.predict([[0]])
> > Out[26]: array([ 0.])
> >
> > In [27]: clf.predict(X)
> > Out[27]: array([ 0., 1., 2.])
> >
> > In [28]: clf.decision_function(X)
> > Out[28]:
> > array([[-0.63212056, -0.98168436, -0.3495638 ],
> > [ 0.63212056, -0. , -0.63212056],
> > [ 0.3495638 , 0.98168436, 0.63212056]])
> >
> >
> -----------------------------------------------------------------------------------------------
> >
> >
> >
> > The decision_function return confusing results. Why [-0.63212056,
> > -0.98168436, -0.3495638 ] corresponds to label 0 ?
> > The encoding of labels seems to be different from the natural orders of
> (0,
> > 1, 2 ...) .
> > After reading the README file of LibSVM, I found the label encoding can
> be
> > obtained by calling svm_get_labels().
> > Where can I find this function wrapper in sklearn? Without that, the
> return
> > results of decision_function() are difficult to interpret.
> > Thanks!
> >
> > --
> > Best Wishes
> > --------------------------------------------
> > Meng Xinfan（蒙新泛）
> > Institute of Computational Linguistics
> > Department of Computer Science & Technology
> > School of Electronic Engineering & Computer Science
> > Peking University
> > Beijing, 100871
> > China
> >
> >
> ------------------------------------------------------------------------------
> > This SF email is sponsosred by:
> > Try Windows Azure free for 90 days Click Here
> > http://p.sf.net/sfu/sfd2d-msazure
> > _______________________________________________
> > Scikit-learn-general mailing list
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
> >
>
>
> ------------------------------------------------------------------------------
> This SF email is sponsosred by:
> Try Windows Azure free for 90 days Click Here
> http://p.sf.net/sfu/sfd2d-msazure
> _______________________________________________
> Scikit-learn-general mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
>
--
Best Wishes
--------------------------------------------
Meng Xinfan（蒙新泛）
Institute of Computational Linguistics
Department of Computer Science & Technology
School of Electronic Engineering & Computer Science
Peking University
Beijing, 100871
China
------------------------------------------------------------------------------
This SF email is sponsosred by:
Try Windows Azure free for 90 days Click Here
http://p.sf.net/sfu/sfd2d-msazure
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
