Orateur
Reda Chhaibi
Description
Consider the basic operation of estimating the spectrum of large covariance matrices.
This estimation has an inherent "large dimensional bias", when one observes a multivariate sample whose size is comparable to the dimension.
Solving this issue amounts to understanding free multiplicative deconvolution.
Our work follows the footsteps of El Karoui, Arizmendi-Tarrago-Vargas and Ledoit-Péché.
After presenting their work, we will discuss the pros and cons of the methods.
Then
1) we will exhibit our own method for computable and statistically consistent estimation.
2) present a cramer-Rao lower bound
This is work in progress. Feedback from the audience will be required.