Estimating large deviation rate functions

Abstract

Establishing a Large Deviation Principle (LDP) proves to be a powerful result for a vast number of stochastic models in many application areas of probability theory. The key object of an LDP is the large deviations rate function, from which probabilistic estimates of rare events can be determined. In order make these results empirically applicable, it would be necessary to estimate the rate function from observations. This is the question we address in this article for the best known and most widely used LDP: Cramér's theorem for random walks. We establish that even when only a narrow LDP holds for Cramér's Theorem, as occurs for heavy-tailed increments, one gets a LDP for estimating the random walk's rate function in the space of convex lower-semicontinuous functions equipped with the Attouch-Wets topology via empirical estimates of the moment generating function. This result may seem surprising as it is saying that for Cramér's theorem, one can quickly form non-parametric estimates of the function that governs the likelihood of rare events.