This talk discusses some attempts on taking a geometric mean of Positive Definite Matrices. This has been of interest in some statistical communities because data sets sometimes naturally come in the form of sequences of random positive definite matrices (diffusion tensor imaging). We attempt to compute the eigenvalue density of the geometric mean of two complex Wishart ensembles. The attempt involves an application of several matrix jacobians, the HCIZ integral formula, an occasional supporting role by the Andreief integration formula along with a cameo appearence of the projective ensemble. We conclude with a very hard integral and a cry for help. Speaker(s): Asad Lodhia (University of Michigan)
Building: | East Hall |
---|---|
Event Type: | Workshop / Seminar |
Tags: | Mathematics |
Source: | Happening @ Michigan from Department of Mathematics, Integrable Systems and Random Matrix Theory Seminar - Department of Mathematics |