OASIS: An interpretable, finite-sample valid alternative to Pearson's <i>X</i>² for scientific discovery.

Additional random train/test splits can be utilized with Bonferroni correction to improve empirical performance, ensuring rejection of highly significant tables at the expense of a higher burden of multiple hypothesis correction (discussed in SI Appendix , section S.6.A ). 4.2.

also in 132,142 other papers