Is it really so difficult to quantify severe uncertainty probabilistically?
Answer-43: This is indeed the question that needs to be asked first and foremost about Info-Gap. And to be sure, this is precisely the question that I am asked most often, mainly by statisticians. I might add that, this question was raised and debated at some length, during a question/answer period following a student’s presentation, in our department, on November 28, 2008. Some of my colleagues go so far as to argue that in the context of a mathematical model, uncertainty should “always” be described by some sort of a probability model. The point is that in those situations where Info-Gap is typically applied to, coming up with a rough probabilistic quantification of the parameter of interest is no more difficult a task than venturing a wild guess about it. The question of course is which is the better practice. Is it preferable (does it more accurately capture the situation?) to “wild-guess” a point estimate of the parameter of interest and conduct a robustness analysis around it —
Related Questions
- Can you quantify the uncertainty in positional measurement introduced by diffuse light and discrete objects within the field of view (field stop) which are not individually detected?
- Is it really so difficult to quantify severe uncertainty probabilistically?
- Is Uncertainty Sidekick Pro difficult to use?