Say you have generated some two dimensional data (a, b) using a Gaussian distribution centered at (0, 0).
If you have a covariance matrix [1, -1/2, -1/2, 1], how do you find potential anomaly points?
For example (1, 0), (0, 1), are they anomaly points?
Or (1, -1)?
I was thinking you do not have the information necessary to solve.