Index: The Book of Statistical ProofsGeneral TheoremsBayesian statisticsProbabilistic modeling ▷ Full probability model

Definition: Consider measured data $y$ and some unknown latent parameters $\theta$. The combination of a generative model for $y$ in terms of the parameters $\theta$ and a prior distribution on $\theta$ in terms of hyperparameters $\lambda$ is called a full probability model $m$:

\[\label{eq:fpm} m: \, y \sim \mathcal{D}_1(\theta), \, \theta \sim \mathcal{D}_2(\lambda) \; .\]
 
Sources:

Metadata: ID: D30 | shortcut: fpm | author: JoramSoch | date: 2020-03-03, 16:16.