[ENH] added parallelization to dominance analysis function

netneurolab / netneurotools

Useful tools from the Network Neuroscience Lab

https://netneurolab.github.io/netneurotools

BSD 3-Clause "New" or "Revised" License

60 stars 32 forks source link

[ENH] added parallelization to dominance analysis function #133

Closed eric2302 closed 8 months ago

eric2302 commented 8 months ago

Model R2s for dominance analysis are now computed in parallel

eric2302 commented 8 months ago

It didn't pass the style check so I'm going to reupload a fancier version 😅

eric2302 commented 8 months ago

I did some tests and it turns out that the version with parallelization produces different array for partial and total dominance sometimes. These are, however, only very minor. Checking with np.isclose() gives a positive result. I've been looking for reasons why this happens, but haven't found any. What do you think @liuzhenqi77 should we merge or not?