On Ranking-based Tests of Independence
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Standard
On Ranking-based Tests of Independence. / Limnios, Myrto; Clémençon, Stephan.
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics. PMLR, 2024. s. 577-585 (Proceedings of Machine Learning Research, Bind 238).Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - On Ranking-based Tests of Independence
AU - Limnios, Myrto
AU - Clémençon, Stephan
N1 - Publisher Copyright: Copyright 2024 by the author(s).
PY - 2024
Y1 - 2024
N2 - In this paper we develop a novel nonparametric framework to test the independence of two random variables X and Y with unknown respective marginals H(dx) and G(dy) and joint distribution F(dxdy), based on Receiver Operating Characteristic (ROC) analysis and bipartite ranking. The rationale behind our approach relies on the fact that, the independence hypothesis H0 is necessarily false as soon as the optimal scoring function related to the pair of distributions (H G, F), obtained from a bipartite ranking algorithm, has a ROC curve that deviates from the main diagonal of the unit square. We consider a wide class of rank statistics encompassing many ways of deviating from the diagonal in the ROC space to build tests of independence. Beyond its great flexibility, this new method has theoretical properties that far surpass those of its competitors. Nonasymptotic bounds for the two types of testing errors are established. From an empirical perspective, the novel procedure we promote in this paper exhibits a remarkable ability to detect small departures, of various types, from the null assumption H0, even in high dimension, as supported by the numerical experiments presented here.
AB - In this paper we develop a novel nonparametric framework to test the independence of two random variables X and Y with unknown respective marginals H(dx) and G(dy) and joint distribution F(dxdy), based on Receiver Operating Characteristic (ROC) analysis and bipartite ranking. The rationale behind our approach relies on the fact that, the independence hypothesis H0 is necessarily false as soon as the optimal scoring function related to the pair of distributions (H G, F), obtained from a bipartite ranking algorithm, has a ROC curve that deviates from the main diagonal of the unit square. We consider a wide class of rank statistics encompassing many ways of deviating from the diagonal in the ROC space to build tests of independence. Beyond its great flexibility, this new method has theoretical properties that far surpass those of its competitors. Nonasymptotic bounds for the two types of testing errors are established. From an empirical perspective, the novel procedure we promote in this paper exhibits a remarkable ability to detect small departures, of various types, from the null assumption H0, even in high dimension, as supported by the numerical experiments presented here.
M3 - Article in proceedings
AN - SCOPUS:85194137713
T3 - Proceedings of Machine Learning Research
SP - 577
EP - 585
BT - Proceedings of The 27th International Conference on Artificial Intelligence and Statistics
PB - PMLR
T2 - 27th International Conference on Artificial Intelligence and Statistics, AISTATS 2024
Y2 - 2 May 2024 through 4 May 2024
ER -
ID: 393771378