LRT-distance that we use to evaluate the performance of our covariance estimates.
slrt(S, trueS)scalar LRT-distance
estimated covariance matrix
true covariance matrix.
Mike Danilov
Note that this is not actually a distance in a sense that slrt(M1,M2) != slrt(M2,M1)
Seber, G.A. (2004) Multivariate observations, Wiley
Danilov, M. (2010). Robust Estimation of Multivariate Scatter under Non-Affine Equivariant Scenarios. Ph.D. thesis, Department of Statistics, University of British Columbia.