Normal Approximation and Concentration for Functions of Sample Covariance Operators

With Vladimir Koltchinskii (Georgia Tech)

Normal Approximation and Concentration for Functions of Sample Covariance Operators

We will discuss several results concerning normal approximation and concentration properties of functions of sample covariance operators, including, in particular, their spectral projection operators
that are of crucial importance in principle component analysis. The problem will be studied in a dimension-free framework with complexity being characterized by the so called “effective rank’’ of the true covariance.

Add to your calendar or Include in your list