| Título: | Combine vector quantization and support vector machine for imbalanced datasets |
| Autores: |
Yu, Ting Debenham, John Jan, Tony Simoff, Simeon |
| Fecha: |
2012-11-08 2006-08 2006-08 |
| Publicador: | Unversidad Nacional de La Plata |
| Fuente: |
 |
| Tipo: |
Objeto de conferencia Objeto de conferencia |
| Tema: |
Ciencias Informáticas Base de Datos |
| Descripción: |
In cases of extremely imbalanced dataset with high dimensions, standard machine learning techniques tend to be overwhelmed by the large classes. This paper rebalances skewed datasets by compressing the majority class. This approach combines Vector Quantization and Support Vector Machine and constructs a new approach, VQ-SVM, to rebalance datasets without significant information loss. Some issues, e.g. distortion and support vectors, have been discussed to address the trade-off between the information loss and undersampling. Experiments compare VQ-SVM and standard SVM on some imbalanced datasets with varied imbalance ratios, and results show that the performance of VQ-SVM is superior to SVM, especially in case of extremely imbalanced large datasets. IFIP International Conference on Artificial Intelligence in Theory and Practice - Integration of AI with other Technologies |
| Idioma: | Inglés |
1 Ajuste de las variables que gobiernan los modelos de comportamiento de HDM-4 para vías no pavimentadas de la región de Antofagasta (Chile) por Rojas Cazaluade, Oscar Orlando | 6 Alternativas para el fraccionamiento de un Parque Industrial por Borrego, Juan Antonio |
2 La Agrimensura como base de investigaciones arqueológicas por Alfageme, Marcelo O.,Papagni, Jorge H. | 7 La propiedad horizontal por López, María Inés |
3 Anteproyecto de saneamiento y loteo por Antolín, Ricardo Gabriel,Williams, Enrique Oliver | 8 Laguna "San Antonio" por Milone, Ricardo Raúl |
4 Balneario Municipal de la ciudad de Diamante por Battistessa, Gustavo J.,Palomeque, José F. | 9 Estudio catastral de tierras en el partido de San Fernando por Miquelarena, Carlos Alberto |
5 | 10 |