Returns the variances of the dirichlet distribution
(α_i * (α_0 - α_i)) / (α_0^2 * (α_0 + 1))
for the i
th element where α_i
is the i
th concentration parameter
and α_0
is the sum of all concentration parameters
Returns the variance of the multinomial distribution
n * p_i * (1 - p_i) for i in 1...k
where n
is the number of trials, p_i
is the i
th probability,
and k
is the total number of probabilities
Returns the covariance matrix of the multivariate normal distribution