Abstract :
[en] The authors demonstrate that the problems encountered with the binning analysis essentially lie in the significance test which is used. They review and analyse most of the existing significance tests and show that the use of those based on randomization processes considerably improves the situation. For the one-dimensional binning analysis, they apply the '2 within 4' randomization test whereas for the two- and the three-dimensional binning analyses they introduce two innovations: the '4 within 16' randomization test and the '8 within 64' randomization test, respectively.
Scopus citations®
without self-citations
1