source: orange/docs/widgets/rst/regression/earth.rst @ 11359:8d54e79aa135

Revision 11359:8d54e79aa135, 2.0 KB checked in by Ales Erjavec <ales.erjavec@…>, 14 months ago (diff)

Cleanup of 'Widget catalog' documentation.

Fixed rst text formating, replaced dead hardcoded reference links (now using
:ref:), etc.

Line 
1.. _Earth:
2
3Earth Learner
4=============
5
6.. image:: ../../../../Orange/OrangeWidgets/icons/Unknown.png
7   :alt: Earth Learner
8   
9Channels
10--------
11
12Inputs:
13   - Data (Table)
14
15Outputs:
16   - Learner
17        The Earth learning algorithm with parameters as specified in the
18         dialog.
19
20   - Predictor
21        Trained regressor
22
23Signal ``Predictor`` sends the regressor only if signal ``Data`` is present.
24
25Description
26-----------
27
28This widget constructs a Earth learning algorithm (an implementation of
29the `MARS - Multivariate Adaptive Regression Splines`_). As all widgets
30for classification and regression, this widget provides a learner and
31classifier/regressor on the output. Learner is a learning algorithm with
32settings as specified by the user. It can be fed into widgets for testing
33learners, for instance Test Learners.
34
35.. _`MARS - Multivariate Adaptive Regression Splines`: http://en.wikipedia.org/wiki/Multivariate_adaptive_regression_splines
36
37.. image:: images/Earth.png
38
39
40.. rst-class:: stamp-list
41
42    1. Learner/Predictor can be given a name under which it will appear
43       in other widgets (say ``Test Learners`` or ``Predictions``).
44
45    2. The ``Max. term degree`` parameter specifies the degree of the
46       terms induced in the forward pass. For instance, if set to ``1``
47       the resulting model will contain only linear terms.
48
49    3. The ``Max. terms`` specifies how many terms can be induces in the
50       forward pass. A special value ``Automatic`` instructs the learner
51       to set the limit automatically based on the dimensionality of the
52       data (``min(200, max(20, 2 * NumberOfAttributes)) + 1``)
53
54    4. The ``Knot penalty`` is used in the pruning pass (hinge function
55       penalty for the GCV calculation)
56
57
58After changing one or more settings, you need to push 5 ``Apply``,
59which will put the new learner on the output and, if the training
60examples are given, construct a new predictor and output it as well.
61
62
63Examples
64--------
65
66Lets use the learner to train a model on a data subset and test it on
67unseen instances.
68
69.. image:: images/Earth-Schema.png
Note: See TracBrowser for help on using the repository browser.