Min-ge Xie - A union of BFF (Bayesian, frequentist and fiducial) inferences by confidence distribution and artificial data sampling
Very interesting talk! What is your response to how CDs perform according to the Martin-Liu validity criterion?
4 months ago by
The "general validity theorem" on Min-ge's slide 30 is a special case of the general validity theory in my work with Chuanhai Liu, e.g., in our 2013 JASA paper.
Also, manipulating the output of this construction as if it were a probability distribution, e.g., marginalization via integration, does not guarantee that validity maintains for inference on the interest parameter. Thinking of the output as a belief/plausibility function, as Chuanhai and I do, provides the mathematical framework for doing marginalization while maintaining validity.
Please login to add an answer/comment or follow this question.