source: orange/docs/reference/rst/SimpleTreeLearner.txt @ 9372:aef193695ea9

Revision 9372:aef193695ea9, 967 bytes checked in by mitar, 2 years ago (diff)

Moved documentation to the separate directory.

Line 
1:obj:`SimpleTreeLearner` is an implementation of regression and classification
2trees. It is faster than :obj:`TreeLearner` at the expense of flexibility.
3It uses gain ratio for classification and MSE for regression.
4
5:obj:`SimpleTreeLearner` was developed for speeding up the construction
6of random forests, but can also be used as a standalone tree.
7
8.. class:: SimpleTreeLearner
9
10    .. attribute:: max_majority
11
12        Maximal proportion of majority class. When this is exceeded,
13        induction stops (only used for classification).
14
15    .. attribute:: min_instances
16
17        Minimal number of instances in leaves. Instance count is weighed.
18
19    .. attribute:: max_depth
20
21        Maximal depth of tree.
22
23    .. attribute:: skip_prob
24
25        At every split an attribute will be skipped with probability ``skip_prob``.
26        Useful for building random forests.
27
28    .. attribute:: random_generator
29       
30        :obj:`Orange.core.RandomGenerator` to use.
Note: See TracBrowser for help on using the repository browser.