Uncertainty quantification

cset2

Statistical inference is composed of two steps, namely estimation, and empirical uncertainty quantification. The latter, which amounts to the construction of a confidence set around the estimated parameter, is crucial for quality assessment and decision making. In high or infinite dimensional models, there is often the paradox that while model adaptive estimation is possible, model adaptive uncertainty quantification is impossible.

 

Last Modification: 01.03.2018 - Contact Person: Dr. Burkhard Thiele