Final published version
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSN › Conference contribution/Paper › peer-review
}
TY - GEN
T1 - Modeling extreme events in spatial domain by copula graphical models
AU - Yu, H.
AU - Choo, Z.
AU - Uy, W.I.T.
AU - Dauwels, J.
AU - Jonathan, P.
PY - 2012
Y1 - 2012
N2 - We propose a new statistical model that captures the conditional dependence among extreme events in a spatial domain. This model may for instance be used to describe catastrophic events such as earthquakes, floods, or hurricanes in certain regions, and in particular to predict extreme values at unmonitored sites. The proposed model is derived as follows. The block maxima at each location are assumed to follow a Generalized Extreme Value (GEV) distribution. Spatial dependence is modeled in two complementary ways. The GEV parameters are coupled through a thin-membrane model, a specific type of Gaussian graphical model often used as smoothness prior. The extreme events, on the other hand, are coupled through a copula Gaussian graphical model with the precision matrix corresponding to a (generalized) thin-membrane model. We then derive inference and interpolation algorithms for the proposed model. The approach is validated on synthetic data as well as real data related to hurricanes in the Gulf of Mexico. Numerical results suggest that it can accurately describe extreme events in spatial domain, and can reliably interpolate extreme values at arbitrary sites. © 2012 ISIF (Intl Society of Information Fusi).
AB - We propose a new statistical model that captures the conditional dependence among extreme events in a spatial domain. This model may for instance be used to describe catastrophic events such as earthquakes, floods, or hurricanes in certain regions, and in particular to predict extreme values at unmonitored sites. The proposed model is derived as follows. The block maxima at each location are assumed to follow a Generalized Extreme Value (GEV) distribution. Spatial dependence is modeled in two complementary ways. The GEV parameters are coupled through a thin-membrane model, a specific type of Gaussian graphical model often used as smoothness prior. The extreme events, on the other hand, are coupled through a copula Gaussian graphical model with the precision matrix corresponding to a (generalized) thin-membrane model. We then derive inference and interpolation algorithms for the proposed model. The approach is validated on synthetic data as well as real data related to hurricanes in the Gulf of Mexico. Numerical results suggest that it can accurately describe extreme events in spatial domain, and can reliably interpolate extreme values at arbitrary sites. © 2012 ISIF (Intl Society of Information Fusi).
KW - Block maxima
KW - Catastrophic event
KW - Conditional dependence
KW - Extreme events
KW - Extreme value
KW - Gaussians
KW - Generalized extreme value
KW - GraphicaL model
KW - Gulf of Mexico
KW - Interpolation algorithms
KW - Numerical results
KW - Precision matrix
KW - Spatial dependence
KW - Spatial domains
KW - Statistical models
KW - Synthetic data
KW - Graphic methods
KW - Hurricanes
KW - Inference engines
KW - Information fusion
KW - Speech recognition
M3 - Conference contribution/Paper
SN - 9781467304177
SP - 1761
EP - 1768
BT - 2012 15th International Conference on Information Fusion
PB - IEEE
ER -