ASAI 2000 - Simposio argentino de inteligencia artificial
Fuente:
Redalyc
Gespeichert in:
| 1. Verfasser: | |
|---|---|
| Format: | Artículo científico |
| Sprache: | en |
| Veröffentlicht: |
Asociación Española para la Inteligencia Artificial
2001
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1876484097814036480 |
|---|---|
| author | Gabriela P. Henning |
| author_facet | Gabriela P. Henning |
| contents | ASAI 2000 - Simposio argentino de inteligencia artificial Gabriela P. Henning Ingeniería Bias Regression Neural Networks Ensemble Learning Variance Decomposition The performance of a single regressor/classifier can be improved by combining the outputs of severalpredictors. This is true provided the combined predictors are accurate and diverse enough, which posses theproblem of generating suitable aggregate members in order to have optimal generalization capabilities. Wepropose here a new method for selecting members of regression/classification ensembles. In particular,using artificial neural networks as learners in a regression context, we show that this method leads to smallaggregates with few but very diverse individual networks. The algorithm is favorably tested against othermethods recently proposed in the literature, producing equal performance on the standard statisticaldatabases used as benchmarks with ensembles that have 75% less members on average 2001 artículo científico 1137-3601 https://www.redalyc.org/articulo.oa?id=92551208 en http://www.redalyc.org/revista.oa?id=925 Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial application/pdf Asociación Española para la Inteligencia Artificial Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial (España) Num.12 Vol.5 |
| format | Artículo científico |
| id | redalyc_92551208 |
| institution | Redalyc |
| language | en |
| publishDate | 2001 |
| publisher | Asociación Española para la Inteligencia Artificial |
| spellingShingle | ASAI 2000 - Simposio argentino de inteligencia artificial Gabriela P. Henning Ingeniería Bias Regression Neural Networks Ensemble Learning Variance Decomposition ASAI 2000 - Simposio argentino de inteligencia artificial Gabriela P. Henning Ingeniería Bias Regression Neural Networks Ensemble Learning Variance Decomposition The performance of a single regressor/classifier can be improved by combining the outputs of severalpredictors. This is true provided the combined predictors are accurate and diverse enough, which posses theproblem of generating suitable aggregate members in order to have optimal generalization capabilities. Wepropose here a new method for selecting members of regression/classification ensembles. In particular,using artificial neural networks as learners in a regression context, we show that this method leads to smallaggregates with few but very diverse individual networks. The algorithm is favorably tested against othermethods recently proposed in the literature, producing equal performance on the standard statisticaldatabases used as benchmarks with ensembles that have 75% less members on average 2001 artículo científico 1137-3601 https://www.redalyc.org/articulo.oa?id=92551208 en http://www.redalyc.org/revista.oa?id=925 Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial application/pdf Asociación Española para la Inteligencia Artificial Inteligencia Artificial. Revista Iberoamericana de Inteligencia Artificial (España) Num.12 Vol.5 |
| title | ASAI 2000 - Simposio argentino de inteligencia artificial |
| topic | Ingeniería Bias Regression Neural Networks Ensemble Learning Variance Decomposition |
| url | https://www.redalyc.org/articulo.oa?id=92551208 |