Fig. 1: Example of an HMM application.
From: Neurophysiological avenues to better conceptualizing adaptive cognition
A single time series (panel a, blue line) consists of a multimodal distribution as visualized by the histogram in panel b. The underlying Markov chain is visualized in panel c. The system cycles through two states, and each state has an observation model, which - in this case - is a univariate random Gaussian distribution characterized by its mean and standard deviation. Inferring the HMM refers to uncovering the states’ parameters and their activations (red line in panel a).
