For reference, it was answered there:
http://stackoverflow.com/questions/31599624/user-defined-svm-kernel-with-scikit-learn
On 07/23/2015 07:33 PM, Vincent Leclère wrote:
Hello everybody,
I'm encountering a trouble fact dealing with sklearn.svm user defined
kernel. Here is a minimal example :
from sklearn.datasets import load_digits from sklearn.svm import SVC
import numpy as np digits = load_digits(2) X, y = digits.data,
digits.target x = X[1,:]
m1 = SVC(kernel='rbf',gamma=1) m1.fit(X, y) print (m1.predict(x)) def
my_kernel(x,y): d = x - y c = d.dot(d.T) return np.exp(-c) m2 =
SVC(kernel=my_kernel) m2.fit(X, y) print (m2.predict(x))
m1 is the standard RBF kernel, x is a data and m1.predict(x) return a
+1/-1 value as intended.
m2 is the same kernel but user implemented, and if m2.fit(X) run,
m2.predict(x) gives an array of +1/-1... Along the same way
m2.predict(X) doesn't work and gives a broadcasting error.
I have also exposed the problem on stackoverflow.
I would be glad for any help.
------------------------------------------------------------------------
Vincent Leclère
------------------------------------------------------------------------------
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
------------------------------------------------------------------------------
_______________________________________________
Scikit-learn-general mailing list
Scikit-learn-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general
