Rank-based concordance for zero-inflated data: New representations, estimators, and sharp bounds

14.11.2025 11:15 – 12:15

RESEARCH INSTITUTE FOR STATISTICS AND INFORMATION SCIENCE SEMINARS

ABSTRACT

Measuring concordance between random variables is central to understanding dependence in data. While classical measures such as Kendall’s tau, Spearman’s rho, or Gini’s gamma are well established for continuous settings, they often break down when data contain ties or excess zeros. This issue is especially relevant for zero-inflated data, where a discrete mass at zero is combined with a continuous component, commonly found in insurance, weather, and biomedical applications. In this talk, I’ll discuss recent developments in adapting rank-based concordance measures to zero-inflated settings. We revisit several familiar measures, clarify their behavior in the presence of zeros, and propose adjusted formulations that remain interpretable and comparable across different data structures. Along the way, we derive sharp bounds for these measures and correct some inconsistencies found in the existing literature. Simulations and real-world case studies illustrate how zero inflation can substantially alter concordance assessment, and how these adjusted measures provide a more reliable picture of association in practice.

(Jointly with Jasper Arends, Guanjie Lyu, Mhamed Mesfioui, and Julien Trufin.)

Lieu

Bâtiment: Uni Mail

Boulevard du Pont-d'Arve 40
1205 Geneva

Room M 5220, 5th floor

entrée libre