Mark Meerschaert, University of Nevada, Reno


Fitting operator stable models to data from finance and hydrology


Many real problems in finance and hydrology involve data sets with power
law tails.  For multivariable data, the power law tail index often
depends on the variable.  Operator stable laws are the central limit
distributions for sums of i.i.d. random vectors with operator norming,
allowing a different tail behavior in each coordinate.  They are the
natural multivariable analogue to stable laws.  In this talk, we will
discuss nonparametric methods for fitting an operator stable
distribution to multivariable heavy tailed data, and we will illustrate
the practical application of these methods with real data sets from
finance and hydrology.