Group Factor Analysis (GFA)

These examples demonstrate Group Factor Analysis, a Bayesian method for discovering shared and group-specific latent structure in multi-group data.

GFA uses Automatic Relevance Determination (ARD) to automatically select the appropriate latent dimensionality during inference.

Start with Simulating from the GFA model to understand the generative model, then proceed through fitting and posterior analysis.

Simulating from the GFA model

Simulating from the GFA model

Fitting GFA to simulated data

Fitting GFA to simulated data

Posterior distributions and sampling

Posterior distributions and sampling

Production workflows

Production workflows