# Generate samples from two multivariate normal densities

the question is:

Generate samples from two multivariate normal densities N(μi, Σi), i = 1, 2, and calculate the Bayes’ optimal discriminant for the cases “Shared, Axis-aligned” and “Shared, Hyperellipsoidal”

Shared, Axis-aligned: where covariance matrix: $S_i = S$, with $s_{ij} = 0$ and $d$ parameters

Shared, Hyperellipsoidal: where covariance matrix: $S_i = S$ and $\frac{d(d+1)}{2}$ parameters

It's a 3rd exercise from book: Machine Learning by Alpaydin, chapter 5, Multivariate Methods

I know I should generate samples that have some variables given and some I should generate. But it's very difficult to understand it from that book. Thank you very much

-