A design criterion for symmetric model discrimination

Description

Experimental design applications for discriminating between models have been hampered by the assumption to know beforehand which model is the true or more adequate one, which is counter to the very aim of the experiment. Previous approaches to alleviate this requirement were either symmetrizations of asymmetric techniques such as compound $T$-optimality, or Bayesian, minimax and sequential approaches. In their talk Harman and Müller presented a novel, genuinely symmetric criterion based on a linearised distance between mean-value surfaces and the newly introduced notion of nominal confidence sets . The computational efficiency of the proposed approach was shown and a Monte-Carlo evaluation of its discrimination performance on the basis of the likelihood-ratio was provided. Additionally Harman and Müller demonstrated the applicability of the new method for a pair of competing models in enzyme kinetics.

Period	07 Aug 2017
Event title	Latest Advances in the Theory and Applications of Design and Analysis of Experiments (17w5007)
Event type	Conference
Location	CanadaShow on map