for X1 = [1,2,4,1]
and X2 = [5,1,2,4]
how do you find the sample means and covariance matrix for the two variables without knowing their distributions ?
I guess you want a 2 by 2 sample variance matrix.
Put the sample variances in the diagonal positions and the sample covariance in the off diagonal spot.
NOW do we use n or n-1 in this case?
Not sure if you want the MLEs (n) or the unbiased estimators (n-1).