|
Hi all, I was reading this paper They say this was the original problem.
Then they say the dual of it is
where $S_j$ is the jth row of the sample covariance matrix. And W is the estimate of the covariance matrix of the Gaussian distribution from from which the samples are drawn. I didn't get how the dual was formed. Is the question clear. If not I will elaborate. Can anyone help please? |
