Consistent Inverse Probability and Possibility Propagation
Dominik Hose, Michael Hanss · 2019
Given a probability distribution of an output quantity of a model, it is generally not possible to infer a unique probability distribution of the uncertain input quantity.In this contribution, it is shown that by reverting back to the coarser framework of possibility theory this problem possesses a conceptually straightforward solution with some powerful properties in the view of imprecise probability descriptions.