Abstract
The problem of assigning a meaningful subjective probability measure on a set of predictive models (inductive hypothesis) is considered. An outperformance construct is suggested as a means of reducing this to the more tractable task of assessing the distribution on a set of realizable events. Further issues on the precise interpretation of subjective probability estimation from the perspective of the decision-oriented forecaster are then considered.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have