Database Reference
In-Depth Information
Fig. 9.4 A prediction from two base classifiers and a single arbiter.
Fig. 9.5 Sample arbiter tree.
which four classifiers
M
1
−
M
4
are generated concurrently.
T
12
and
T
34
are
the training sets generated by the rule selection from which arbiters are
produced.
A
12
and
A
34
are the two arbiters. Similarly,
T
14
and
A
14
(root
arbiter) are generated and the arbiter tree is completed.
There are several schemes for arbiter trees; each is characterized by a
different selection rule. Here are three versions of selection rules:
•
Only instances with classifications that disagree are chosen (group 1).
•
Like group 1 defined above, plus instances where their classifications
agree but are incorrect (group 2).
•
Like groups 1 and 2 defined above, plus instances that have the same
correct classifications (group 3).
Of the two versions of arbitration rules that have been implemented, each
corresponds to the selection rule used for generating the training data at
that level:
•
For selection rules 1 and 2, a final classification is made by a majority
vote of the classifications of the two lower levels and the arbiter's own
classification, with preference given to the latter.
Search WWH ::
Custom Search