Proto Distributions

• Distributions
• Occur in Bayesian

Continuous Spaces

• Proto PDF
• $p(x|\theta )=\frac{1}{{\int }_{{\mathbb{R}}^k}{p}_0(x|\theta )dx}{p}_0(x|\theta )$
• $p_0$ gives the shape of the PDF
• $\frac{1}{{\int }_{{\mathbb{R}}^k}{p}_0(x|\theta )dx}$ is a normalization so it integrates to 1
• 
• Most of the time we dont know a distribution but only its proto distribution. This is actually enough sometimes

Discrete Spaces

• Proto PMF