Why Gaussian mixture model uses Expectation maximization instead of Gradient descent?
What other models uses Expectation maximization to find best optimal parameters instead of using gradient descent?
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
Not all the parameters (e.g., the assignment parameters) for a Gaussian mixture model are smoothly differentiable, thus can not be fit with gradient descent.
Other use cases for the expectation–maximization (EM) algorithm are:
- Clustering
- Latent variable estimation
- Missing data estimation
answered Jul 6, 2020 at 14:35
-
$\begingroup$ Thanks ,Could you please explain on what is meant by smoothly differentiable ? , i know that we have to find the best value for the mean vector ,covariance matrix using Expectation maximization $\endgroup$– starCommented Jul 6, 2020 at 16:21