This question is motivated by this post. Then the log multivariate normal density looks like this:

\begin{equation} \implies \log(1)+ \frac{p}{2}\log(2\pi)+ \frac{1}{2}\log(|\Sigma^X_j+\Sigma^Y_{k}|^{-1})- \frac{1}{2}(Z_i-\mu^X_j)^T(\Sigma^X_j+\Sigma^Y_k)^{-1}(Z_i-\mu^X_j) \end{equation}

Now, One way is to use EM to update the estimates, but in M-step, one could encounter taking derivative of $\log(det(\Sigma^X_j+\Sigma^Y_{k})^{-1})$ or derivative of above equation respect to $\Sigma^X_j$ or $\Sigma^Y_k$, how can this be done? Many thanks in advance.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.