Changeset 10712:428fb1432e9e in orange


Ignore:
Timestamp:
04/03/12 10:24:36 (2 years ago)
Author:
anze <anze.staric@…>
Branch:
default
Message:

Fixes to documentation of Orange.tuning.

Files:
3 edited

Legend:

Unmodified
Added
Removed
  • Orange/tuning/__init__.py

    r10711 r10712  
    1 """  
    2  
    3     
    4 """ 
    5  
    61import Orange.core 
    72import Orange.classification 
  • docs/reference/rst/Orange.tuning.rst

    r10711 r10712  
     1.. py:currentmodule:: Orange.classification.majority 
     2 
    13############################### 
    2 Optimization (``optimization``) 
     4Tuning (``tuning``) 
    35############################### 
    46 
    5 .. automodule:: Orange.optimization 
     7.. automodule:: Orange.tuning 
    68 
    79.. index:: tuning 
     
    1820 
    1921Two classes support tuning parameters. 
    20 :obj:`Orange.optimization.Tune1Parameter` for fitting a single parameter and 
    21 :obj:`Orange.optimization.TuneMParameters` fitting multiple parameters at once, 
     22:obj:`~Tune1Parameter` for fitting a single parameter and 
     23:obj:`~TuneMParameters` fitting multiple parameters at once, 
    2224trying all possible combinations. When called with data and, optionally, id 
    2325of meta attribute with weights, they find the optimal setting of arguments 
    2426using cross validation. The classes can also be used as ordinary learning 
    2527algorithms - they are in fact derived from 
    26 :obj:`Orange.classification.Learner`. 
     28:obj:`~Orange.classification.Learner`. 
    2729 
    28 Both classes have a common parent, :obj:`Orange.optimization.TuneParameters`, 
     30Both classes have a common parent, :obj:`~TuneParameters`, 
    2931and a few common attributes. 
    3032 
     
    7678The script first divides the data into training and testing subsets. It trains 
    7779a naive Bayesian classifier and than wraps it into 
    78 :obj:`Orange.optimization.ThresholdClassifiers` with thresholds of .2, .5 and 
     80:obj:`~ThresholdClassifiers` with thresholds of .2, .5 and 
    7981.8. The three models are tested on the left-out data, and we compute the 
    8082confusion matrices from the results. The printout:: 
  • docs/reference/rst/index.rst

    r10581 r10712  
    3232   Orange.network 
    3333 
    34    Orange.optimization 
     34   Orange.tuning 
    3535    
    3636   Orange.projection 
Note: See TracChangeset for help on using the changeset viewer.