Agnostic insurability of model classes
Abstract
Motivated by problems in insurance, our task is to predict finite upper bounds on a future draw from an unknown distribution over the set of natural numbers. We can only use past observations generated independently and identically distributed according to . While is unknown, it is known to belong to a given collection of probability distributions on the natural numbers. The support of the distributions may be unbounded, and the prediction game goes on for \emph{infinitely} many draws. We are allowed to make observations without predicting upper bounds for some time. But we must, with probability 1, start and then continue to predict upper bounds after a finite time irrespective of which governs the data. If it is possible, without knowledge of and for any prescribed confidence however close to 1, to come up with a sequence of upper bounds that is never violated over an infinite time window with confidence at least as big as prescribed, we say the model class is \emph{insurable}. We completely characterize the insurability of any class of distributions over natural numbers by means of a condition on how the neighborhoods of distributions in should be, one that is both necessary and sufficient.
Cite
@article{arxiv.1212.3866,
title = {Agnostic insurability of model classes},
author = {Narayana Santhanam and Venkat Anantharam},
journal= {arXiv preprint arXiv:1212.3866},
year = {2013}
}