80 search results - page 1 / 16 » A Dynamic Classifier Selection Method to Build Ensembles usi... |
SBRN
2006
IEEE
13 years 10 months ago
2006
IEEE
WAIM
2004
Springer
13 years 10 months ago
2004
Springer
Abstract. Ensemble methods can achieve excellent performance relying on member classifiers’ accuracy and diversity. We conduct an empirical study of the relationship of ensemble...
ICMLA
2008
13 years 6 months ago
2008
The test set accuracy for ensembles of classifiers selected based on single measures of accuracy and diversity as well as combinations of such measures is investigated. It is foun...
IJCAI
2003
13 years 6 months ago
2003
Ensemble methods like bagging and boosting that combine the decisions of multiple hypotheses are some of the strongest existing machine learning methods. The diversity of the memb...
DMIN
2006
13 years 6 months ago
2006
- An ideal ensemble is composed of base classifiers that perform well and that have minimal overlap in their errors. Eliminating classifiers from an ensemble based on a criterion t...