Changeset 10282:68cbf304dfdc in orange


Ignore:
Timestamp:
02/16/12 15:19:26 (2 years ago)
Author:
anzeh <anze.staric@…>
Branch:
default
rebase_source:
662e0b77e1af8d62b25a3f2ce682cd6770d37df2
Message:

Fixed documentation for scoring.CA.

Files:
2 edited

Legend:

Unmodified
Added
Removed
  • Orange/evaluation/scoring.py

    r10280 r10282  
    357357 
    358358class CA(list): 
    359     """Computation of CA from different types of test_results""" 
     359    """ 
     360    Compute percentage of matches between predicted and actual class. 
     361 
     362    :param test_results: :obj:`~Orange.evaluation.testing.ExperimentResults` 
     363                         or list of :obj:`ConfusionMatrix`. 
     364    :param report_se: include standard error in result. 
     365    :param ignore_weights: ignore instance weights. 
     366    :rtype: list of scores, one for each learner. 
     367 
     368    Standard errors are estimated from deviation of CAs across folds (if 
     369    test_results were produced by cross_validation) or approximated under 
     370    the assumption of normal distribution otherwise. 
     371    """ 
    360372    CONFUSION_MATRIX = 0 
    361373    CONFUSION_MATRIX_LIST = 1 
     
    366378                          "unweighted": "ignore_weights"}) 
    367379    def __init__(self, test_results, report_se = False, ignore_weights=False): 
    368         """Return percentage of matches between predicted and actual class. 
    369  
    370         :param test_results: :obj:`~Orange.evaluation.testing.ExperimentResults` 
    371                              or :obj:`ConfusionMatrix`. 
    372         :param report_se: include standard error in result. 
    373         :param ignore_weights: ignore instance weights. 
    374         :rtype: list of scores, one for each learner. 
    375  
    376         Standard errors are estimated from deviation of CAs across folds (if 
    377         test_results were produced by cross_validation) or approximated under 
    378         the assumption of normal distribution otherwise. 
    379         """ 
    380380        super(CA, self).__init__() 
    381381        self.report_se = report_se 
  • docs/reference/rst/Orange.evaluation.scoring.rst

    r10258 r10282  
    2727be passed as well. 
    2828 
    29 .. autosingleton:: CA 
    30 .. autoclass:: CAClass 
    31     :members: 
     29.. autoclass:: CA 
    3230.. autofunction:: sens 
    3331.. autofunction:: spec 
     
    7472Regression 
    7573========== 
    76  
    77 General Measure of Quality 
    78 ========================== 
    7974 
    8075Several alternative measures, as given below, can be used to evaluate 
Note: See TracChangeset for help on using the changeset viewer.