Abstract
It is well understood that the covariance matrix plays an important role for both theoretical and applied statistical problems. Furthermore, for practitioners it is of utmost importance to be able to find and use a covariance matrix that adequately describes dependence in the data, for example among repeated measurements. A parsimonious statistical modelling often yields structured/patterned covariance matrices which are also justified in many real-life problems.
In this talk, spectral properties and inverses of some patterned matrices used in statistics are demonstrated. In particular, block matrices involving Kronecker structure are of interest.