Open rafaqz opened 3 months ago
Maybe useful, but not licensed: https://github.com/lbreston/CCS
Looks useful, @rafaqz! From a brief glance at their supplementary information, this looks pretty straight forward to implement.
One parameter in their method is T
, the fraction of pairwise distances considered. They say "Rougher signals require a larger fraction of pairwise distances to be considered because their ranks are less tightly coupled". To determine roughness, they use a measure involving the standard deviation. We should be maximally flexible, so we should implement a dispatch-based method for allowing not only standard deviation but other measures of "data spread" when calculating T
.
Another one....
https://www.nature.com/articles/s41598-021-98864-2 https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10336384/#RSOS221590C29