TY - GEN
T1 - Ensemble Kalman filter regularization using leave-one-out data cross-validation
AU - Rayo Schiappacasse, Lautaro Jerónimo
AU - Hoteit, Ibrahim
N1 - KAUST Repository Item: Exported on 2020-10-01
PY - 2012/9/27
Y1 - 2012/9/27
N2 - In this work, the classical leave-one-out cross-validation method for selecting a regularization parameter for the Tikhonov problem is implemented within the EnKF framework. Following the original concept, the regularization parameter is selected such that it minimizes the predictive error. Some ideas about the implementation, suitability and conceptual interest of the method are discussed. Finally, what will be called the data cross-validation regularized EnKF (dCVr-EnKF) is implemented in a 2D 2-phase synthetic oil reservoir experiment and the results analyzed.
AB - In this work, the classical leave-one-out cross-validation method for selecting a regularization parameter for the Tikhonov problem is implemented within the EnKF framework. Following the original concept, the regularization parameter is selected such that it minimizes the predictive error. Some ideas about the implementation, suitability and conceptual interest of the method are discussed. Finally, what will be called the data cross-validation regularized EnKF (dCVr-EnKF) is implemented in a 2D 2-phase synthetic oil reservoir experiment and the results analyzed.
UR - http://hdl.handle.net/10754/552767
UR - http://scitation.aip.org/content/aip/proceeding/aipcp/10.1063/1.4756379
UR - http://www.scopus.com/inward/record.url?scp=84883107431&partnerID=8YFLogxK
U2 - 10.1063/1.4756379
DO - 10.1063/1.4756379
M3 - Conference contribution
SN - 9780735410916
SP - 1247
EP - 1250
BT - International Conference of Numerical Analysis and Applied Mathematics, ICNAAM 2012
PB - AIP Publishing
ER -