In the Documentation: SVMClassifier Documentation there are class SimpleCV.MachineLearning.SVMClassifier.SVMClassifier(featureExtractors, properties=None)
mSVMProperties = {
'KernelType':'RBF', #default is a RBF Kernel
'SVMType':'NU', #default is C
'nu':None, # NU for SVM NU
'c':None, #C for SVM C - the slack variable
'degree':None, #degree for poly kernels - defaults to 3
'coef':None, #coef for Poly/Sigmoid defaults to 0
'gamma':None, #kernel param for poly/rbf/sigma - default is 1/#samples
}
...
self.mKernelType = {
'RBF':orange.SVMLearner.RBF, #Radial basis kernel
'Linear':orange.SVMLearner.Linear, #Linear basis kernel
'Poly':orange.SVMLearner.Polynomial, #Polynomial kernel
'Sigmoid':orange.SVMLearner.Sigmoid #Sigmoid Kernel
}
...
self.mSVMType = {
'NU':orange.SVMLearner.Nu_SVC,
'C':orange.SVMLearner.C_SVC
}
How to set KernelType? ex:to 'Linear', i've tried with this code
classifier = SVMClassifier(feature_extractors,{'KernelType':'Linear'})
there was an error
ERROR:
Traceback (most recent call last):
File "ntitrasi.py", line 56, in <module>
classifier = SVMClassifier(feature_extractors,{'KernelType':'Linear'})
File "C:\Python27\lib\site-packages\simplecv-1.3-py2.7.egg\SimpleCV\MachineLea
rning\SVMClassifier.py", line 70, in __init__
self._parameterizeKernel()
File "C:\Python27\lib\site-packages\simplecv-1.3-py2.7.egg\SimpleCV\MachineLea
rning\SVMClassifier.py", line 90, in _parameterizeKernel
self.mSVMPrototype.svm_type = self.mSVMType[self.mSVMProperties["SVMType"]]
KeyError: 'SVMType'
and i've tried this one too
classifier = SVMClassifier(feature_extractors,{'KernelType':'Linear','SVMType':'C'})
and there was an error too:
ERROR:
Traceback (most recent call last):
File "ntitrasi.py", line 56, in <module>
classifier = SVMClassifier(feature_extractors,{'KernelType':'Linear','SVMTyp
e':'C'})
File "C:\Python27\lib\site-packages\simplecv-1.3-py2.7.egg\SimpleCV\MachineLea
rning\SVMClassifier.py", line 70, in __init__
self._parameterizeKernel()
File "C:\Python27\lib\site-packages\simplecv-1.3-py2.7.egg\SimpleCV\MachineLea
rning\SVMClassifier.py", line 92, in _parameterizeKernel
if(self.mSVMProperties["nu"] is not None):
KeyError: 'nu'
Did you try to include nu
in the parameter list?:
classifier = SVMClassifier(feature_extractors,{'KernelType':'Linear','SVMType':'C','nu':None})?