wrapper for no ICI case

[rmflight]

for those who want just a fast version of cor(..., method = "kendall"), should we have a wrapper that works almost the same way as the base correlation function?

something like

fast_kt(data, use = c("everything", "complete", "pairwise.complete")

And then it just does the basic kendall-tau, but fast?

[hunter-moseley]

Yes.

On Tue, Jul 11, 2023, 10:01 AM Robert M Flight @.***> wrote:

for those who want just a fast version of cor(..., method = "kendall"), should we have a wrapper that works almost the same way as the base correlation function?

something like

fast_kt(data, use = c("everything", "complete", "pairwise.complete")

And then it just does the basic kendall-tau, but fast?

— Reply to this email directly, view it on GitHub https://github.com/MoseleyBioinformaticsLab/ICIKendallTau/issues/11, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADEP7B43LOGA76OJDSYPUNDXPVMCRANCNFSM6AAAAAA2GB43YA . You are receiving this because you are subscribed to this thread.Message ID: @.***>

[rmflight]

We discussed this in today's lab meeting, but I think the design should be:

• Take a matrix or two variables to calculate on, as well as what to use (everything, complete, pairwise complete).
• If there are NAs, figure out how to do the pairwise comparisons.
• Including where we can return NA correlation values in the resultant correlations.
• Pass only clean data with no missing values to the underlying ici_kt function, so we won't need to modify any code in the C++.