For a covariance matrix, I am aware of the precision matrix, the covariance matrix inverse. What's the analog for that for a cross covariance matrix, i.e. $E[XY^{\top}]-E[X]E[Y^{\top}]$ for two random vectors $X$ and $Y$ of different length? Would it be the pseudo inverse of the cross-covariance matrix?
Asked
Active
Viewed 10 times
