Modelling heavy tail data with bayesian nonparametric mixtures
Luis E. Nieto-Barajas · 2026
Read Modelling heavy tail data with bayesian nonparametric mixtures by Luis E. Nieto-Barajas (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-07228.txt.
Description
Heavy-tailed data, from insurance losses to network traffic, is often handled by throwing away the bulk of the data and modeling only extremes. Luis E. Nieto-Barajas instead models the whole spread with Bayesian nonparametric mixtures, keeping body and tail.
The mixtures use shifted gamma-gamma components with a Poisson-Dirichlet mixing spread. One setting in each component tracks tail heaviness, so its posterior spread assesses how heavy each tail is. An efficient Markov chain Monte Carlo method with adaptive steps carries out the inference, illustrated on simulated and real data sets. Stats experts get a unified model that wastes no data while still answering the tail question.
Catalog record
| CREATOR | Luis E. Nieto-Barajas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 38 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07228