Aggregation algorithms for regression : A comparison with boosting and SVM techniques
Classi cation and regression ensembles sho w generalization capabilities that outperform those of single predictors. We present here a further ev aluation of tw o algorithms for ensemble construction recently proposed by us. In particular, we compare them with Boosting and Support Vector Machine tec...
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Objeto de conferencia |
| Lenguaje: | Inglés |
| Publicado: |
2003
|
| Materias: | |
| Acceso en línea: | http://sedici.unlp.edu.ar/handle/10915/22869 |
| Aporte de: |
| Sumario: | Classi cation and regression ensembles sho w generalization capabilities that outperform those of single predictors. We present here a further ev aluation of tw o algorithms for ensemble construction recently proposed by us. In particular, we compare them with Boosting and Support Vector Machine tec hniques, which are the newest and most sophisticated methods to treat classi cation and regression problems. We sho w that our comparatively simpler algorithms are very competitive with these tec hniques, showing even a sensible improvement in performance in some of the standard statistical databases used as benchmarks. |
|---|