Latent Structure & Model Comparison

This week addresses models with hidden structure — components you cannot observe directly but whose existence shapes the data you can see. Mixture models decompose a dataset into overlapping populations. Latent variable models formalise the idea that unobserved quantities drive observed correlations. Model comparison then asks: given two competing explanations, which one should you prefer?

Mixture Models

TBD

Latent Variable Models

TBD

Comparing And Selecting Models

TBD