|
Abstract
|
|
|---|---|
| The naive Bayes is a competitive classifier that makes strong conditional independence assumptions. Its accuracy can be improved by relaxing these assumptions. One classifier which does that is the semi-naive Bayes. The state-of-the-art algorithm for learning a semi-naive Bayes from data is the backward sequential elimination and joining (BSEJ) algorithm. We extend BSEJ with a second step which removes some of its unwarranted independence assumptions. Our classifier outperforms BSEJ and five other Bayesian network classifiers on a set of benchmark databases, although the difference in performance is not statistically significant. | |
|
International
|
Si |
|
|
|
|
Book Edition
|
|
|
Book Publishing
|
Springer |
|
ISBN
|
978-3-642-40642-3 |
|
Series
|
|
|
Book title
|
Advances in Artificial Intelligence, Lecture Notes in Artificial Intelligence 8109 |
|
From page
|
159 |
|
To page
|
167 |