Exploring the Bivariate Normal Distribution

A random vector \((X_1, X_2)\) has a bivariate normal distribution with mean \(\boldsymbol{\mu} = (0, 0)\) and variance–covariance matrix

\[ \Sigma = \begin{pmatrix} \sigma_1^2 & \sigma_{12} \\ \sigma_{12} & \sigma_2^2 \end{pmatrix} \]

if its density is

\[ f(x_1, x_2) = \frac{1}{2\pi \sqrt{\det \Sigma}} \exp\!\left( -\tfrac{1}{2}\, \mathbf{x}^\top \Sigma^{-1} \mathbf{x} \right). \]

Type a matrix below to see its density. The diagonal entries are the variances, and the off-diagonal entry is the covariance; because \(\Sigma\) is symmetric, editing one off-diagonal entry updates the other. \(\Sigma\) must be positive definite: both variances positive and \(\sigma_1^2 \sigma_2^2 - \sigma_{12}^2 > 0\). Drag the plot to rotate it and scroll to zoom.

\(\Sigma =\)

Marginal distributions

Each coordinate on its own is normal, with mean 0 and variance given by the matching diagonal entry of \(\Sigma\): \(X_1 \sim N(0, \sigma_1^2)\) and \(X_2 \sim N(0, \sigma_2^2)\). The covariance \(\sigma_{12}\) has no effect on the marginals; it only changes how \(X_1\) and \(X_2\) vary together.