search

fmorenopino / HeterogeneousHMM

Discrete, Gaussian, and Heterogenous HMM models full implemented in Python. Missing data, Model Selection Criteria (AIC/BIC), and Semi-Supervised training supported. Easily extendable with other types of probablistic models.
https://pyhhmm.readthedocs.io/en/latest/
Apache License 2.0
74 stars 14 forks source link

n_g_emissions and n_d_emissions parameters #2

Closed milliemince closed 2 years ago

milliemince commented 2 years ago

Hello - want to use a heterogeneous HMM for a project I'm working on and have a question about the parameters. Say my feature space has 8 Gaussian features and 4 discrete features. I'm assuming then that n_g_emissions = 8 and n_d_emissions = 4... but I'm wondering how the model can tell which features are the Gaussian ones and which are the discrete ones. Is it as simple as putting the Gaussian features before the discrete ones in the input feature matrix? Or does your code somehow recognize which features are Gaussian and which are discrete?

Any clarification on this would be highly appreciated. Thank you!

semese commented 2 years ago

Hi,

In order for the model to deal with this you have to arrange your features in the feature matrix such that the first n_g_emissions columns are the Gaussian features, and the last n_d_emissions columns are the discrete features.

I hope this helped. Do not hesitate to contact us if you encounter further problems.