Skip to content

Permutation inference for Canonical Correlation Analysis

License

Notifications You must be signed in to change notification settings

andersonwinkler/PermCCA

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

31 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PermCCA

This repository contains a reference implementation for statistically valid permutation tests with canonical correlation analysis.

The paper reference is:

The implementation of both Huh-Jhun and Theil are slightly more involved in the code than in the paper. Both implementations are expected to lead to identical results. Here in the repository, the implementation of Huh-Jhun uses null(Z'), which appears to be numerically more stable in Matlab/Octave. For Theil, there are more options for selecting observations to be dropped (via selection matrix, numerical indices, logical indices, or random selection). These can be helpful in practice. In the paper only the generalised case of using a selection matrix is shown.

About

Permutation inference for Canonical Correlation Analysis

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages