Ignore:
Timestamp:
05/23/12 16:02:44 (2 years ago)
Author:
markotoplak
Branch:
default
Message:

Logistic regression with LibLinear: added an explicit bias parameter, some documentation changes.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • Orange/classification/logreg.py

    r10684 r10883  
    10191019 
    10201020    def __init__(self, solver_type=L2R_LR, C=1.0, eps=0.01, normalization=True, 
    1021                 **kwargs): 
     1021            bias=-1.0, **kwargs): 
    10221022        """ 
    10231023        :param solver_type: One of the following class constants:  
    10241024            ``L2_LR`` or``L2_LR_DUAL`` 
    10251025         
    1026         :param C: Regularization parameter (default 1.0) 
     1026        :param C: Regularization parameter (default 1.0). Higher values of C mean  
     1027            less regularization (C is a coefficient for the loss function). 
    10271028        :type C: float 
    10281029         
     
    10331034            (default True) 
    10341035        :type normalization: bool 
     1036 
     1037        :param bias: If positive, use it as a bias (default -1). 
     1038        :type bias: float 
    10351039         
    10361040        """ 
     
    10391043        self.eps = eps 
    10401044        self.normalization = normalization 
    1041  
    1042         for name, valu in kwargs.items(): 
     1045        self.bias = bias 
     1046 
     1047        for name, value in kwargs.items(): 
    10431048            setattr(self, name, value) 
    10441049 
Note: See TracChangeset for help on using the changeset viewer.