CM3 processes model the spatio-temporal dependence structure for extreme values of functional data fields and M4 processes for discrete data fields. Given a dataset, this software estimates the length of the tail dependence, the number of patterns of extreme values and the patterns with their relative frequencies of occurrence. It yields a full model. The simulation mode enables you to fit the best parameters before studying real examples. Five demo files are included, namely with an application to electricity spot prices.
General routines contained inside: Partitioning Around Medoids (PAM), non parametric unit-Frechet standardization, classical runs estimator of the extremal index.
Thomas Meinguet (2020). MatEx - Matlab extremes (https://www.mathworks.com/matlabcentral/fileexchange/26539-matex-matlab-extremes), MATLAB Central File Exchange. Retrieved .
Demo5: application to electricity spot prices.
Last version of Demo4.m
Corrections in Demo3.m and Demo4.m