Changeset 7306:edfa72151fb1 in orange
 Timestamp:
 02/03/11 11:51:26 (3 years ago)
 Branch:
 default
 Convert:
 3746ac6d54359e357cefcd9fb2d9a30767f19b6f
 File:

 1 edited
Legend:
 Unmodified
 Added
 Removed

orange/Orange/classification/svm/__init__.py
r7294 r7306 12 12 13 13 .. note:: On some datasets SVM can perform very badly. It is a known fact that 14 SVM 'scan be very sensitive to the proper choice of the parameters.15 If you are having problems with learner's accuracy try scaling the14 SVM can be very sensitive to the proper choice of the parameters. 15 If you are having problems with the learner's accuracy try scaling the 16 16 data and using different parameters or choose an easier approach 17 and use the `SVMLearnerEasy` class which does this automatically18 (it is similar to the easy.pyscript in the LibSVM distribution).17 and use the :obj:`SVMLearnerEasy` class which does this automatically 18 (it is similar to the `svmeasy.py`_ script in the LibSVM distribution). 19 19 20 20 .. autoclass:: Orange.classification.svm.SVMLearner … … 154 154 155 155 class SVMLearner(_SVMLearner): 156 """:param svm_type: defines the type of SVM (can be C_SVC, Nu_SVC (default), OneClass, Epsilon_SVR, Nu_SVR) 157 :type svm_type: SVMLearner.SVMType 158 :param kernel_type: defines the type of a kernel to use for learning 159 (can be kernels.RBF (default), kernels.Linear, kernels.Polynomial, 160 kernels.Sigmoid, kernels.Custom) 161 :type kernel_type: classification.kernels.Kernel 162 :param degree: kernel parameter (for Polynomial) (default 3) 163 :type degree: int 164 :param gamma: kernel parameter (Polynomial/RBF/Sigmoid) 165 (default 1/number_of_examples) 166 :type gamma: float 167 :param coef0: kernel parameter (Polynomial/Sigmoid) (default 0) 168 :type coef0: int 169 :param kernelFunc: function that will be called if `kernel_type` is 170 `Custom`. It must accept two `data.Instance` arguments and 171 return a float (the distance between the examples). 172 :type kernelFunc: callable function 173 :param C: C parameter for C_SVC, Epsilon_SVR, Nu_SVR 174 :type C: float 175 :param nu: Nu parameter for Nu_SVC, Nu_SVR and OneClass (default 0.5) 176 :type nu: float 177 :param p: epsilon in lossfunction for Epsilon_SVR 178 :type p: float 179 :param cache_size: cache memory size in MB (default 100) 180 :type cache_size: int 181 :param eps: tolerance of termination criterion (default 0.001) 182 :type eps: float 183 :param probability: determines if a probability model should be build 184 (default False) 185 :type probability: bool 186 :param shrinking: determines whether to use shrinking heuristics 187 (default True) 188 :type shrinking: bool 189 :param weights: a list of class weights 190 :type weights: list 191 192 """ 156 193 __new__ = _orange__new__(_SVMLearner) 157 194 … … 167 204 cache_size=200, eps=0.001, normalization=True, 168 205 weight=[], **kwargs): 169 """:param svm_type: defines the type of SVM (can be C_SVC,170 Nu_SVC (default), OneClass, Epsilon_SVR, Nu_SVR)171 :type svm_type: SVMLearner.SVMType172 :param kernel_type: defines the type of a kernel to use for learning173 (can be kernels.RBF (default), kernels.Linear, kernels.Polynomial,174 kernels.Sigmoid, kernels.Custom)175 :type kernel_type: classification.kernels.Kernel176 :param degree: kernel parameter (for Polynomial) (default 3)177 :type degree: int178 :param gamma: kernel parameter (Polynomial/RBF/Sigmoid)179 (default 1/number_of_examples)180 :type gamma: float181 :param coef0: kernel parameter (Polynomial/Sigmoid) (default 0)182 :type coef0: int183 :param kernelFunc: function that will be called if `kernel_type` is184 `Custom`. It must accept two `data.Instance` arguments and185 return a float (the distance between the examples).186 :type kernelFunc: callable function187 :param C: C parameter for C_SVC, Epsilon_SVR, Nu_SVR188 :type C: float189 :param nu: Nu parameter for Nu_SVC, Nu_SVR and OneClass (default 0.5)190 :type nu: float191 :param p: epsilon in lossfunction for Epsilon_SVR192 :type p: float193 :param cache_size: cache memory size in MB (default 100)194 :type cache_size: int195 :param eps: tolerance of termination criterion (default 0.001)196 :type eps: float197 :param probability: determines if a probability model should be build198 (default False)199 :type probability: bool200 :param shrinking: determines whether to use shrinking heuristics201 (default True)202 :type shrinking: bool203 :param weights: a list of class weights204 :type weights: list205 206 """207 206 self.svm_type = SVMLearner.Nu_SVC 208 207 self.kernel_type = kernel_type
Note: See TracChangeset
for help on using the changeset viewer.