General framework for class-specific feature selection
Commonly, when a feature selection algorithm is applied, a single feature subset is selected for all the classes, but this subset could be inadequate for some classes. Class-specific feature selection allows selecting a possible different feature subset for each class. However, all the class-specifi...
| Autores: | , , |
|---|---|
| Tipo de recurso: | artículo |
| Estado: | Versión aceptada para publicación |
| Fecha de publicación: | 2011 |
| País: | México |
| Institución: | Instituto Nacional de Astrofísica, Óptica y Electrónica |
| Repositorio: | Repositorio Institucional del INAOE |
| Idioma: | inglés |
| OAI Identifier: | oai:inaoe.repositorioinstitucional.mx:1009/1593 |
| Acceso en línea: | http://inaoe.repositorioinstitucional.mx/jspui/handle/1009/1593 |
| Access Level: | acceso abierto |
| Palabra clave: | info:eu-repo/classification/Class-specific feature selection/Class-specific feature selection info:eu-repo/classification/Feature selection/Feature selection info:eu-repo/classification/Supervised classification/Supervised classification info:eu-repo/classification/Classifier ensemble/Classifier ensemble info:eu-repo/classification/cti/1 info:eu-repo/classification/cti/12 info:eu-repo/classification/cti/1203 |
| Sumario: | Commonly, when a feature selection algorithm is applied, a single feature subset is selected for all the classes, but this subset could be inadequate for some classes. Class-specific feature selection allows selecting a possible different feature subset for each class. However, all the class-specific feature selection algorithms have been proposed for a particular classifier, which reduce their applicability. In this paper, a general framework for using any traditional feature selector for doing class-specific feature selection, which allows using any classifier, is proposed. Experimental results and a comparison against traditional feature selectors showing the suitability of the proposed framework are included. |
|---|