Webinar, February 4: Epistemic confidence, the Dutch Book and relevant subsets

On February 4 at 16.00, Yudi Pawitan will give a webinar: Epistemic confidence, the Dutch Book and relevant subsets.

Abstract as follows:

Epistemic confidence is defined as the sense of confidence in an observed confidence interval. This epistemic property is accepted in the Bayesian philosophy, but unavailable — or even denied — in orthodox frequentist inference. My goal is to describe a nonBayesian way to establish an epistemic confidence. In financial markets, including the betting market, the Dutch Book is known as arbitrage or risk-free profitable trade. A numerical confidence is deemed epistemic if its use as a betting price is protected from the Dutch Book by an external agent. Theoretically, to construct the Dutch Book, the agent must exploit unused information available in any relevant subset. Pawitan and Lee (2020) showed that confidence is an extended likelihood, and the likelihood principle states that the likelihood contains all the information in the data, hence leaving no relevant subset. This implies that confidence associated with the full likelihood is protected from the Dutch Book, hence epistemic. I will describe some theory involving relevant subsets and illustrate it with examples.

(This talk is based on joint work with Youngjo Lee and Hangbin Lee from Seoul National University, South Korea.)

The meeting will take place in Zoom, and the number to use is 66273335562 .