Talk from Archives

On The Modeling of Covariance Matrices: A Generalized Fisher Transformation (joint work with Peter Reinhard Hansen (Univ. North Carolina, Chapel Hill))

05.11.2018 16:45 - 17:45

For the modeling of covariance matrices, the literature has proposed a variety of methods to enforce the positive (semi) definiteness. In this paper, we propose a novel approach with attractive properties and a wide range of applications. The methods is based on the matrix logarithmic transformation of the correlation matrix, and we present evidence that the resulting off-diagonal elements constitute the generalization of Fisher's Z-transformation of a single correlation.

Personal website of Ilya Archakov

Location:
HS 7 OMP1 (#1.303)