Description
It is a type of hidden Markov model (HMM) in which the hidden states are Gaussian mixtures.
They are typically used to model systems that can switch between two different states, such as a light switch that can be either on or off, or a gene that can be either expressed or not expressed. However, they can also be used to model systems with more than two states.
Switch Binary SGMs are trained on data that consists of observations of the system’s state at different points in time.