Ignore:
Timestamp:
05/03/12 12:40:00 (2 years ago)
Author:
Ales Erjavec <ales.erjavec@…>
Branch:
default
Message:

Added a note to 'confusion_matrices' function for the return value if
the 'class_index' is not specified.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • Orange/evaluation/scoring.py

    r10784 r10857  
    671671    :param test_results: test results 
    672672    :param class_index: index of class value for which the confusion matrices 
    673                         are to be computed. 
     673        are to be computed (by default unspecified - see note below). 
    674674    :param ignore_weights: ignore instance weights. 
    675     :params cutoff: cutoff for probability 
    676  
    677     :rtype: list of :obj:`ConfusionMatrix` 
     675    :param cutoff: cutoff for probability 
     676 
     677    :rtype: list of :obj:`ConfusionMatrix` or list-of-list-of-lists (see 
     678        note below) 
     679 
     680    .. note:: If `class_index` is not specified and `test_results` 
     681        contain predictions for multi-class problem, then the return 
     682        value is a list of 2D tables (list-of-lists) of all class 
     683        value pairwise misclassifications.  
     684 
    678685    """ 
    679686    tfpns = [ConfusionMatrix() for _ in range(test_results.number_of_learners)] 
Note: See TracChangeset for help on using the changeset viewer.