Linear-Mapping based Variational Ensemble Kalman Filter

Linjie Wen, Jinglai Li

We propose a linear-mapping based variational Ensemble Kalman filter for problems with generic observation models. Specifically, the proposed method is formulated as to construct a linear mapping from the prior ensemble to the posterior one, and the linear mapping is computed by minimizing the Kullback-Leibler divergence between the transformed distribution by the linear mapping and the actual posterior.

Knowledge Graph

arrow_drop_up

Comments

Sign up or login to leave a comment