hhg-test

Heller-Heller-Gorfine multivariate test of association

Getting Started

inversions.c must be compiled from C source code into a MEX file.

Prerequisites

This code was tested in MATLAB R2017b. A MATLAB-compatible C compiler is required, as is the MATLAB Statistics and Machine Learning Toolbox.

Usage

HHGPermutationTest takes five input arguments:

1. X: The first input data matrix. Rows are assumed to represent samples, and columns are assumed to represent dimensions.
2. Y: The second input data matrix. Y must contain the same number of samples as X.
3. nperm: The number of permutations to perform. The default value is 100.
4. maxN: The maximum sample size for which the full distance matrices will be held in memory. If the sample size exceeds maxN, the distance matrices will be computed incrementally and stored on disk.

Once computation is completed, HHGPermutationTest returns up to three output arguments:

1. p: The p-value of the permutation test. If p < 1/nperm, a value of 0 is returned.
2. t: The value of the HHG test statistic.
3. pstat: The values of the HHG test statistic for each permutation.

References

1. Heller, R., Heller, Y., & Gorfine, M. (2012). "A consistent multivariate test of association based on ranks of distances." Biometrika, 100(2), 503-510. (Link)

License

This project is licensed under the MIT License.