Laplace distribution (Double exponential)¶

This distribution is robust to outliers, and puts more weight to values closer to 0 than a Gaussian distribution. This property is useful to encourage sparsity in a model.

\[X \sim Lap(\mu, b)\]

PDF:¶

\[Lap(x| \mu, b) \triangleq \frac{1}{2b}\exp {( - \frac{|x - \mu |}{b} )}\]

Here:

  • \(\mu\) is a location parameter

  • \(b\) is a scale parameters

Moments¶

Mean¶

\(E[X] = \mu\)

Variance¶

\(var[X] = 2b^2\)