Agnostic insurability of model classes

MaRDI QIDQ5744840zbMATH OpenFDO

Authors Narayana P. Santhanam, Venkat Anantharam

Publication date 19 February 2016

Full work available at URL https://arxiv.org/abs/1212.3866, http://jmlr.csail.mit.edu/papers/v16/santhanam15a.html

zbMATH Keywords

insurance of probability distributions over countable setsnon-parametric approachesprediction of quantiles of distributionsuniversal compression

Mathematics Subject Classification ID

Applications of statistics to actuarial sciences and financial mathematics (62P05)

Abstract: Motivated by problems in insurance, our task is to predict finite upper bounds on a future draw from an unknown distribution $p$ over the set of natural numbers. We can only use past observations generated independently and identically distributed according to $p$. While $p$ is unknown, it is known to belong to a given collection $calP$ of probability distributions on the natural numbers. The support of the distributions $pincalP$ 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 $pincalP$ governs the data. If it is possible, without knowledge of $p$ 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 $calP$ is emph{insurable}. We completely characterize the insurability of any class $calP$ of distributions over natural numbers by means of a condition on how the neighborhoods of distributions in $calP$ should be, one that is both necessary and sufficient.

Recommendations

• Valid model-free prediction of future insurance claims
• scientific article; zbMATH DE number 4113817
• Prediction in a Poisson cluster model
• scientific article; zbMATH DE number 4174199
• On the risk of credibility premium rules

Cited in

(3)

• scientific article; zbMATH DE number 44317 (Why is no real title available?)
• scientific article; zbMATH DE number 7625180 (Why is no real title available?)
• Non-uniform consistency of online learning with random sampling

This page was built for publication: Agnostic insurability of model classes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5744840)