|
Abstract
|
|
|---|---|
| In multidimensional classi?cation the goal is to assign an instance to a set of different classes. This task is normally addressed either by de?ning a compound class variable with all the possible combinations of classes (label power-set methods, LPMs) or by building independent classi?ers for each class (binary-relevance methods, BRMs). However, LPMs do not scale well and BRMs ignore the dependency relations between classes. We introduce a method for chaining binary Bayesian classi?ers that combines the strengths of classi- ?er chains and Bayesian networks for multidimensional classi?cation. The method consists of two phases. In the ?rst phase, a Bayesian network (BN) that represents the dependency relations between the class variables is learned from data. In the second phase, several chain classi?ers are built, such that the order of the class variables in the chain is consistent with the class BN. At the end we combine the results of the different generated orders. Our method considers the dependencies between class variables and takes advantage of the conditional independence relations to build simpli- ?ed models. We perform experiments with a chain of na¿?ve Bayes classi?ers on different benchmark multidimensional datasets and show that our approach outperforms other state-of-the-art methods. | |
|
International
|
Si |
|
Congress
|
International Joint Conference on Artificial Intelligence |
|
|
960 |
|
Place
|
Barcelona |
|
Reviewers
|
Si |
|
ISBN/ISSN
|
978-1-57735-515-1 |
|
|
|
|
Start Date
|
16/07/2011 |
|
End Date
|
22/07/2011 |
|
From page
|
2192 |
|
To page
|
2197 |
|
|
Proceedings of the Twenty-Second International Joint Conference on Artificial Intelligence |