DECISION FOREST GENERATION
An exemplary method of establishing a decision tree includes determining an effectiveness indicator for each of a plurality of input features. The effectiveness indicators each correspond to a split on the corresponding input feature. One of the input features is selected as a split variable for the split. The selection is made using a weighted random selection that is weighted according to the determined effectiveness indicators.
1 . A method of establishing a decision tree, comprising the steps of:
determining an effectiveness indicator for each of a plurality of input features, the effectiveness indicators each corresponding to a split on the corresponding input feature; and
selecting one of the input features as a split variable for the split using a weighted random selection that is weighted according to the determined effectiveness indicators.
2 . The method of claim 1 , wherein each effectiveness indicator corresponds to a probability that the split on the input feature will yield a useful determination within the decision tree.
3 . The method of claim 1 , wherein
the weighted random selection includes increasing a likelihood of selecting a first one of the input features over a second one of the input features; and
the effectiveness indicator of the first one of the input features is higher than the effectiveness indicator of the second one of the input features.
4 . The method of claim 3 , comprising
weighting the weighted random selection in proportion to the effectiveness indicators.
5 . The method of claim 1 , comprising
limiting which of the input features are candidates for the selecting by selecting only from among the input features that have a effectiveness indicator that exceeds a threshold.
6 . The method of claim 1 , comprising
selectively altering an influence that the effectiveness indicators have on the weighted random selection.
7 . The method of claim 6 , comprising
applying an influencing factor to the effectiveness indicators in a manner that reduces any differences between the effectiveness indicators.
8 . The method of claim 1 , wherein the plurality of input features comprises all input features that are utilized within a random decision forest that includes the established decision tree.
9 . The method of claim 1 , comprising performing the determining and the selecting for at least one split in each of a plurality of decision trees within a random decision forest.
10 . The method of claim 1 , comprising performing the determining and the selecting for each of a plurality of splits in the decision tree.
11 . A device that establishes a decision tree, comprising:
a processor and data storage associated with the processor, the processor being configured to use at least one of instructions or information in the data storage to
determine an effectiveness indicator for each of a plurality of input features, the effectiveness indicators each corresponding to a split on the corresponding input feature and
select one of the input features as a split variable for the split using a weighted random selection that is weighted according to the determined effectiveness indicators.
12 . The device of claim 11 , wherein each effectiveness indicator corresponds to a probability that the split on the input feature will yield a useful determination within the decision tree.
13 . The device of claim 11 , wherein
the weighted random selection includes an increased likelihood of selecting a first one of the input features over a second one of the input features; and
the effectiveness indicator of the first one of the input features is higher than the effectiveness indicator of the second one of the input features.
14 . The device of claim 13 , wherein the processor is configured to
weight the weighted random selection in proportion to the effectiveness indicators.
15 . The device of claim 11 , wherein the processor is configured to
limit which of the input features are candidates to select by selecting only from among the input features that have a effectiveness indicator that exceeds a threshold.
16 . The device of claim 11 , wherein the processor is configured to
selectively alter an influence that the effectiveness indicators have on the weighted random selection.
17 . The device of claim 16 , wherein the processor is configured to
apply an influencing factor to the effectiveness indicators in a manner that reduces any differences between the effectiveness indicators.
18 . The device of claim 11 , wherein the plurality of input features comprises all input features that are utilized within a random decision forest that includes the established decision tree.
19 . The device of claim 11 , wherein the processor is configured to determine the effectiveness indicators and select one of the input features for at least one split in each of a plurality of decision trees within a random decision forest.
20 . The device of claim 11 , wherein the processor is configured to determine the effectiveness indicators and select one of the input features for each of a plurality of splits in the decision tree.