Measuring association with Wasserstein distances. (English) Zbl 07594079

Summary: Let \(\pi \in \Pi (\mu,\nu)\) be a coupling between two probability measures \(\mu\) and \(\nu\) on a Polish space. In this article we propose and study a class of nonparametric measures of association between \(\mu\) and \(\nu \), which we call Wasserstein correlation coefficients. These coefficients are based on the Wasserstein distance between \(\nu\) and the disintegration \(\pi_{x_1}\) of \(\pi\) with respect to the first coordinate. We also establish basic statistical properties of this new class of measures: we develop a statistical theory for strongly consistent estimators and determine their convergence rate in the case of compactly supported measures \(\mu\) and \(\nu \). Throughout our analysis we make use of the so-called adapted/bicausal Wasserstein distance, in particular we rely on results established in [Backhoff, Bartl, Beiglböck, Wiesel. Estimating processes in adapted Wasserstein distance. 2020]. Our approach applies to probability laws on general Polish spaces.


62Gxx Nonparametric inference
62Hxx Multivariate analysis
90Cxx Mathematical programming


Full Text: DOI arXiv Link


[1] Backhoff, J., Bartl, D., Beiglböck, M. and Wiesel, J. \[(2021+)\]. Estimating processes in adapted Wasserstein distance. Ann. Appl. Probab. Avaible at arXiv:2002.07261. · Zbl 1492.60104
[2] Backhoff-Veraguas, J., Bartl, D., Beiglböck, M. and Eder, M. (2020). Adapted Wasserstein distances and stability in mathematical finance. Finance Stoch. 24 601-632. 10.1007/s00780-020-00426-3 · Zbl 1440.91036
[3] Blum, J.R., Kiefer, J. and Rosenblatt, M. (1961). Distribution free tests of independence based on the sample distribution function. Ann. Math. Stat. 32 485-498. 10.1214/aoms/1177705055 · Zbl 0139.36301
[4] Cao, S. and Bickel, P.J. (2020). Correlations with tailored extremal properties. arXiv preprint. Available at arXiv:2008.10177.
[5] Chatterjee, S. (2020). A new coefficient of correlation. J. Amer. Statist. Assoc. 1-21.
[6] Cifarelli, D.M. and Regazzini, E. (2017). On the centennial anniversary of Gini’s theory of statistical relations. Metron 75 227-242. 10.1007/s40300-017-0108-0 · Zbl 1384.62166
[7] Gini, C. (1915). Nuovi contributi alla teoria delle relazioni statistiche. Premiate Officine Grafiche di Carlo Ferrari.
[8] Deb, N., Ghosal, P. and Sen, B. (2020). Measuring association on topological spaces using kernels and geometric graphs. arXiv preprint. Available at arXiv:2010.01768.
[9] Dette, H., Siburg, K.F. and Stoimenov, P.A. (2013). A copula-based non-parametric measure of regression dependence. Scand. J. Stat. 40 21-41. 10.1111/j.1467-9469.2011.00767.x · Zbl 1259.62050
[10] Fournier, N. and Guillin, A. (2015). On the rate of convergence in Wasserstein distance of the empirical measure. Probab. Theory Related Fields 162 707-738. 10.1007/s00440-014-0583-7 · Zbl 1325.60042
[11] Friedman, J.H. and Rafsky, L.C. (1983). Graph-theoretic measures of multivariate association and prediction. Ann. Statist. 11 377-391. · Zbl 0528.62052
[12] Gamboa, F., Klein, T. and Lagnoux, A. (2018). Sensitivity analysis based on Cramér-von Mises distance. SIAM/ASA J. Uncertain. Quantificat. 6 522-548. 10.1137/15M1025621 · Zbl 1392.62091
[13] Geenens, G. and Lafaye de Micheaux, P. (2020). The Hellinger correlation. J. Amer. Statist. Assoc. 1-15.
[14] Giacomo Nies, T., Staudt, T. and Munk, A. (2021). Transport dependency: Optimal transport based dependency measures. arXiv preprint. Available at arXiv:2105.02073.
[15] Griessenberger, F., Junker, R.R. and Trutschnig, W. (2021). On a multivariate copula-based dependence measure and its estimation. arXiv preprint. Available at arXiv:2109.12883. · Zbl 07524973
[16] Junker, R.R., Griessenberger, F. and Trutschnig, W. (2021). Estimating scale-invariant directed dependence of bivariate distributions. Comput. Statist. Data Anal. 153 107058. 10.1016/j.csda.2020.107058 · Zbl 1510.62253
[17] Ke, C. and Yin, X. (2020). Expected conditional characteristic function-based measures for testing independence. J. Amer. Statist. Assoc. 115 985-996. 10.1080/01621459.2019.1604364 · Zbl 1445.62091
[18] Lassalle, R. (2018). Causal transport plans and their Monge-Kantorovich problems. Stoch. Anal. Appl. 36 452-484. 10.1080/07362994.2017.1422747 · Zbl 1385.93087
[19] Linfoot, E.H. (1957). An informational measure of correlation. Inf. Control 1 85-89. · Zbl 0080.36001
[20] Lyons, R. (2013). Distance covariance in metric spaces. Ann. Probab. 41 3284-3305. 10.1214/12-AOP803 · Zbl 1292.62087
[21] Mordant, G. and Segers, J. (2022). Measuring dependence between random vectors via optimal transport. J. Multivariate Anal. 189 104912. 10.1016/j.jmva.2021.104912 · Zbl 1493.62359
[22] Móri, T.F. and Székely, G.J. (2019). Four simple axioms of dependence measures. Metrika 82 1-16. 10.1007/s00184-018-0670-3 · Zbl 1417.62159
[23] Móri, T.F. and Székely, G.J. (2020). The Earth mover’s correlation. Ann. Univ. Sci. Budapest. Sect. Comput. 50 249-268. 10.1007/s10701-018-0208-6 · Zbl 1474.62204
[24] Ozair, S., Lynch, C., Bengio, Y., van den Oord, A., Levine, S. and Sermanet, P. (2019). Wasserstein dependency measure for representation learning. arXiv preprint. Available at arXiv:1903.11780.
[25] Pflug, G.C. (2009). Version-independence and nested distributions in multistage stochastic optimization. SIAM J. Optim. 20 1406-1420. 10.1137/080718401 · Zbl 1198.90307
[26] Pflug, G.C. and Pichler, A. (2012). A distance for multistage stochastic optimization models. SIAM J. Optim. 22 1-23. 10.1137/110825054 · Zbl 1262.90118
[27] Puccetti, G. (2019). Measuring linear correlation between random vectors. Available at SSRN 3116066.
[28] Rényi, A. (1959). On measures of dependence. Acta Math. Acad. Sci. Hung. 10 441-451. 10.1007/BF02024507 · Zbl 0091.14403
[29] Rosenblatt, M. (1975). A quadratic measure of deviation of two-dimensional density estimates and a test of independence. Ann. Statist. 3 1-14. · Zbl 0325.62030 · doi:10.1214/aos/1176342996
[30] Scarsini, M. (1984). On measures of concordance. Stochastica 8 201-218. · Zbl 0582.62047
[31] Schweizer, B. and Wolff, E.F. (1981). On nonparametric measures of dependence for random variables. Ann. Statist. 9 879-885. · Zbl 0468.62012
[32] Shi, H., Drton, M. and Han, F. (2020). Distribution-free consistent independence tests via center-outward ranks and signs. J. Amer. Statist. Assoc. 1-16.
[33] Shi, H., Drton, M. and Han, F. (2020). On the power of Chatterjee rank correlation. arXiv preprint. Available at arXiv:2008.11619.
[34] Sommerfeld, M. and Munk, A. (2018). Inference for empirical Wasserstein distances on finite spaces. J. R. Stat. Soc. Ser. B. Stat. Methodol. 80 219-238. 10.1111/rssb.12236 · Zbl 1380.62121
[35] Székely, G.J., Rizzo, M.L. and Bakirov, N.K. (2007). Measuring and testing dependence by correlation of distances. Ann. Statist. 35 2769-2794. 10.1214/009053607000000505 · Zbl 1129.62059
[36] Trutschnig, W. (2011). On a strong metric on the space of copulas and its induced dependence measure. J. Math. Anal. Appl. 384 690-705. 10.1016/j.jmaa.2011.06.013 · Zbl 1252.46019
[37] Weed, J. and Bach, F. (2019). Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance. Bernoulli 25 2620-2648. 10.3150/18-BEJ1065 · Zbl 1428.62099
[38] Wiesel, J.C.W. (2022). Supplement to “Measuring association with Wasserstein distances.” 10.3150/21-BEJ1438SUPP
[39] Xiao, Y. and Wang, W.Y. (2019). Disentangled representation learning with Wasserstein total correlation. arXiv preprint. Available at arXiv:1912.12818.
[40] Zhang, K. (2019). BET on independence. J. Amer. Statist. Assoc. 114 1620-1637. 10.1080/01621459.2018 · Zbl 1428.62181
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.