NonGaussian Models
Gonzalo R. Arce · 2004
The Gaussian distribution model is widely accepted in signal processing practice. There are, however, applications where the underlying random processes do not follow Gaussian statistics. In order to model nonGaussian processes, a wide variety of distributions with heavier-than-Gaussian tails have been proposed as viable alternatives. This chapter reviews several of these approaches and focuses on two distribution families, namely the class of generalized Gaussian distributions and the class of stable distributions.