The Matrix Cookbook

Learn / The Matrix Cookbook

07 - Multivariate Distributions

Advanced formula selection, assumption checking, and exact application for Multivariate Distributions, book pages 37-39.

Learning path

0%

0 of 2 sections marked complete · about 25 minutes

Learning objectives

What you will be able to explain

  • Distribution supports
  • Distribution roles
  • Multivariate Gaussian density structure
  • Dirichlet and multinomial
  • Heavy-tail limiting behavior
  • Wishart versus inverse Wishart

Section 01

Distribution supports to Multivariate Gaussian density structure

01

Distribution supports

Support is the first guard against applying a density to an invalid object.

Guided checkpoint

Match each distribution to its natural support or object.

Source: Secs. 7.1-7.9, pp. 37-39

02

Distribution roles

The chapter collects density forms and normalization constants for these distinct domains.

Guided checkpoint

Match the family to its defining role.

Source: Secs. 7.1-7.8, pp. 37-38

03

Multivariate Gaussian density structure

The determinant controls volume and the precision controls Mahalanobis geometry.

Guided checkpoint

For xRdx\in R^d and positive-definite SigmaSigma.

Source: Secs. 7.3 and 7.5, p. 37

Section 02

Dirichlet and multinomial to Wishart versus inverse Wishart

01

Dirichlet and multinomial

The simplex constraint and count constraint are complementary.

Guided checkpoint

Judge the conjugate pair.

Source: Secs. 7.2 and 7.6, p. 37

02

Heavy-tail limiting behavior

The t family approaches Gaussian behavior as degrees of freedom increase.

Guided checkpoint

What happens to multivariate Student's t as degrees of freedom grow?

Source: Sec. 7.7, p. 37

03

Wishart versus inverse Wishart

Inversion changes determinants, trace terms, and normalization through the transformation Jacobian.

Guided checkpoint

Check each distinction.

Source: Secs. 7.8-7.9, pp. 38-39

Knowledge check

Turn understanding into recall.

The quiz now follows the same concepts in scored form. You can return to this lesson from the quiz whenever a gap appears.