Manifolds of Differentiable Densities
Abstract
We develop a family of infinite-dimensional (non-parametric) manifolds of probability measures. The latter are defined on underlying Banach spaces, and have densities of class with respect to appropriate reference measures. The case , in which the manifolds are modelled on Fr\'{e}chet spaces, is included. The manifolds admit the Fisher-Rao metric and, unusually for the non-parametric setting, Amari's -covariant derivatives for all . By construction, they are -embedded submanifolds of particular manifolds of finite measures. The statistical manifolds are dually () flat, and admit mixture and exponential representations as charts. Their curvatures with respect to the -covariant derivatives are derived. The likelihood function associated with a finite sample is a continuous function on each of the manifolds, and the -divergences are of class .
Cite
@article{arxiv.1608.03979,
title = {Manifolds of Differentiable Densities},
author = {Nigel J. Newton},
journal= {arXiv preprint arXiv:1608.03979},
year = {2018}
}
Comments
Version 3: 27 pages. Introduction expanded to discuss applications. Concluding Remarks section added. Improved definition of tangent space (space of signed measures). Discussion of Bayesian data fusion expanded. Discussion of normal charts for the $\alpha$ covariant derivatives added. New references added. No change to results. To appear in ESAIM:Probability and Statistics www.esaim-ps.org