On pairwise interaction multivariate Pareto models
It is shown that no multivariate Pareto model, other than the Hüsler–Reiss family, can have the structure of a pairwise interaction model.
Publications
You can also find my articles on my Google Scholar profile.
Learning extremal graphical structures in high dimensions
With S. Engelke and S. Volgushev.
In submission, 2022+ (arXiv)
This paper proposes a methodology that provably learns extremal graphical models in settings where the dimension is allowed to grow exponentially in the effective sample size. Along the way, we prove a sub-exponential concentration inequality for the empirical version of the extremal variogram, an object of intrest in multivariate and high-dimensional extreme value theory.
A new family of smooth copulas with arbitrarily irregular densities
With R. Zimmerman.
In submission, 2022+ (arXiv)
In this paper, we construct a class of copulas that have densities and uniformly bounded high order partial derivatives, but whose densities can be arbitrarily irregular. In particular, we identify a copula density that is finite but nowhere bounded.
Rank-based estimation under asymptotic dependence and independence, with applications to spatial extremes
With S. Engelke and S. Volgushev.
Annals of Statistics 49(5), 2552-2576 (arXiv, published, supplement)
This paper provides theoretical tools and a new methodology to fit flexible bivariate and spatial tail dependence models that include both asymptotic dependence and independence.