Extreme value analysis in Julia

Extremes.jl provides exhaustive high-performance functions for the analysis of extreme values in Julia. In particular, methods for block maxima and threshold exceedances models are implemented, such as:

  • Parameter estimation by probability weighted moments, maximum likelihood and Bayesian procedure.
  • Stationary and non-stationary models.
  • Diagnostic plots for assessing model accuracy.
  • Return level estimation.

The package is largely based on the books of Coles (2001), An Introduction to Statistical Modeling of Extreme Values, and of Beirlant et al. (2004) Statistics of Extremes: Theory and Applications. The Tutorial section illustrates the package functionalities by reproducing many results obtained by Coles (2001).