Stochastic models, information theory, and Lie groups. II: Analytic methods and modern applications (Q638915)

The subject of stochastic processes may be conditionally divided into several branches. The branch originating from Shannon's mathematical theory of communication is covered in numerous engineering textbooks. A second branch from information theory, due to Wiener, is concerned with filtering of noisy data and extracting signals. The third branch originated from the field of mathematical statistics in which researchers like Fisher, Cramér and Rao developed concepts in statistical estimation. Volume 2 is a more formal and more advanced presentation that builds on the basics covered in Volume 1 (2009) (for a review see Zbl 1182.60001). It is composed of three parts. Part 1 begins with a detailed treatment of Lie groups including elementary algebraic, differential geometric, and functional analytic properties. Classical variational calculus techniques are reviewed, and the coordinate-free extensions of these concepts to Lie groups (in the form of the Euler-Poincaré equation) are derived and used in examples. In addition, the basic concepts of group representation theory are reviewed along with the concepts of convolution of functions and Fourier expansions on Lie groups. Connections with multivariate statistical analysis and integral geometry are also explored. Part 2 of Volume 2 is concerned with the connections between information theory and group theory. An extension of the de Bruijn inequality to the context of Lie groups is examined. Classical communication theory problems are reviewed, and information inequalities that have parallels in group theory are explained. Geometric and algebraic problems in coding theory are also examined. A number of connections to problems in engineering and biology are provided.