Facultade de Fisioterapia

Approximating the null distribution of a class of statistics for testing independence

Jiménez Gamero, María Dolores; Alba-Fernández, M. V.; Ariza-López, F. J.
Abstract:
A class of tests for testing independence whose test statistic is an -norm of the difference between the joint empirical characteristic function and the product of the marginal empirical characteristic functions associated with a sample is considered. Since the null distribution of these test statistics is unknown, some approximations are investigated. Specifically, the permutation, bootstrap and weighted bootstrap estimators are examined. All of them provide consistent estimators. A simulation study analyzes the performance of these approximations for small and moderate sample sizes. An application to a real data set for testing independence between positional errors in spatial data is included.
Year:
2019
Type of Publication:
Article
Keywords:
Testing for independenceEmpirical characteristic functionPermutationBootstrapWeighted bootstrapConsistency
Journal:
Journal of Computational and Applied Mathematics
Volume:
354
Pages:
131-143
Month:
July
Note:
Q1 49/252 h-index 1,632 (JCR2018)
Comments:
MTM2014–55966–P and MTM2017–89422–P
DOI:
https://doi.org/10.1016/j.cam.2018.03.011
Hits: 1523