Multivariate Gaussian/Normal Mixture Distribution/Model/Process
- is a type of Process
- is a distribution/model/process that is both:
Bivariate Trimodal Model (Example)
Example Covariances
samples from a Gaussian mixture model. In the example below, there are three components. From left to right:
- the first component has an isotropic covariance matrix, meaning it has the same amount of variance in each direction
- the second has a diagonal covariance matrix, meaning it can control the variance separately along each axis-aligned direction. This example has more variance along the 𝑥2 axis than along the 𝑥1 axis
- the third component has a full-rank covariance matrix, allowing it to control the variance separately along with an arbitrary basis of directions
