! Work in progress !
This is a Julia package that extends Manifolds.jl
to statistical manifolds in information geometry. Essentially, a Euclidean manifold where the local metric is given by the Fisher information matrix.
Example ways to create a manifold
M = FisherMetricManifold(2,p -> fim(Normal(p...)))
M = FisherMetricManifold(Beta)
Once created, one can calculate the distance between two points, p
and q
, in addition to calculating the shortest geodesic connecting p
and q
(amoung several other functions provided by Manifolds.jl
).
using Manifolds
using FisherMetrics
using Distributions
M = FisherMetricManifold(Normal)
p = [1.0,2.0]
q = [5.0,4.0]
B = default_basis(M,p)
dist = distance(M,p,q)
geo = shortest_geodesic(M,p,q)
- Provide test functions.
- Provide documentation.
- Create a more robust boundary value problem solver to compute geodesics (unstable for points that are far away).