Scipy Multivariate Normal Change Parameters

Question

I have a 2-d multivariate Gaussian defined as

from scipy.stats import multivariate_normal
...
gauss_0_rv = multivariate_normal(means[0], covariances[0])

In my EM algorithm, I am calculating new mean and covarience in each iteration. Is there a way to assert new parameters to my gaussian RV, or do I have to define a new one and discard/delete the old distribution?

Answer 1

Yes. Just call multivariate_normal.pdf(x, mean, cov).