turkmath.org

The set of probability distributions constitute a statistical model as a manifold. By means of this model the relationship between the geometric structure of the manifold and statistical estimation can be analyzed. The Fisher information metric plays an important role in information geometry. The pair (M,g) where M is a flat affine manifold and g is a Kaehler affine metric is called a Kaehler affine manifold. A connection, which is also flat, dual to flat affine connection is defined by the Kaehler affine metric. In addition, one can form a family of dual connections that are not flat. The concept of a Kaehler affine metric is naturally arises in information geometry. In this talk, the close relationship between information and affine geometry will be explained. This talk is a part of the TÜBİTAK 1001 113F296 Project called Harmonic maps, affine geometry and their applications to information geometry.