Changeset 10335:18f3ac9e1ec6 in orange


Ignore:
Timestamp:
02/22/12 12:07:50 (2 years ago)
Author:
Lan Zagar <lan.zagar@…>
Branch:
default
Message:

Expanded description of MultitargetVariance.

File:
1 edited

Legend:

Unmodified
Added
Removed
  • Orange/multitarget/tree.py

    r10092 r10335  
    6262class MultitargetVariance(Orange.feature.scoring.Score): 
    6363    """ 
    64     A multi-target score that ranks features based on the variance of the 
    65     subsets. A weighted distance can be used to compute the variance. 
     64    A multi-target score that ranks features based on the average class 
     65    variance of the subsets. 
     66 
     67    To compute it, a prototype has to be defined for each subset. Here, it 
     68    is just the mean vector of class variables. Then the sum of squared 
     69    distances to the prototypes is computed in each subset. The final score 
     70    is obtained as the average of subset variances (weighted, to account for 
     71    subset sizes). 
     72     
     73    Weights can be passed to the constructor to normalize classes with values 
     74    of different magnitudes or to increase the importance of some classes. In 
     75    this case, class values are first scaled according to the given weights. 
    6676    """ 
    6777 
    6878    def __init__(self, weights=None): 
    6979        """ 
    70         :param weights: Weights of the features used when computing distances. 
    71                         If None, all weights are set to 1. 
     80        :param weights: Weights of the class variables used when computing 
     81                        distances. If None, all weights are set to 1. 
    7282        :type weigts: list 
    7383        """ 
Note: See TracChangeset for help on using the changeset viewer.