 |
| Titel |
Flood Classification Using Support Vector Machines |
| VerfasserIn |
Lieke A. Melsen, Paul J. J. Torfs, Claudia C. Brauer |
| Konferenz |
EGU General Assembly 2013
|
| Medientyp |
Artikel
|
| Sprache |
Englisch
|
| Digitales Dokument |
PDF |
| Erschienen |
In: GRA - Volume 15 (2013) |
| Datensatznummer |
250076590
|
|
|
|
|
|
| Zusammenfassung |
| Lowland floods are in general considered to be less extreme than mountainous floods. In
order to investigate this, seven lowland floods in the Netherlands were selected and compared
to mountainous floods from the study of Marchi et al. (2010). Both a 2D and 3D approach of
the statistical two-group classification method support vector machines (Cortes and Vapnik,
1995) were used to find a statistical difference between the two flood types. Support vector
machines were able to draw a decision plane between the two flood types, misclassifying one
out of seven lowland floods, and one out of 67 mountainous floods. The main difference
between the two flood types can be found in the runoff coefficient (with lowland floods
having a lower runoff coefficient than mountainous floods), the cumulative precipitation
causing the flood (which was lower for lowland floods), and, obviously, the relief
ratio.
Support vector machines have proved to be useful for flood classification and might be
applicable in future classification studies.
References
Cortes, C., and V. Vapnik. "Support-Vector Networks." Machine Learning 20: (1995)
273-297.
Marchi, L., M. Borga, E. Preciso, and E. Gaume. "Characterisation of selected extreme flash
floods in Europe and implications for flood risk management." Journal of Hydrology 394:
(2010) 118-133. |
|
|
|
|
|
|