As pointed out in #68, some estimators do not center data internally prior to estimating the covariance matrix. This isn't an issue when using cvCovEst() since it centers the data internally. It prevents the use of the packages estimator functions from being used as standalone estimators, however. The data should be centered by each estimator instead, even if it is slightly less computationally efficient.
As pointed out in #68, some estimators do not center data internally prior to estimating the covariance matrix. This isn't an issue when using
cvCovEst()
since it centers the data internally. It prevents the use of the packages estimator functions from being used as standalone estimators, however. The data should be centered by each estimator instead, even if it is slightly less computationally efficient.