romain-lopez / scvi-reproducibility Goto Github PK

I think the NMI and ARI measures are inverse in the plots(results/clustering) for RETINA dataset.
Because in RETINA-scVI.ipynb you have :
NMI = 0.836 , ARI = 0.615
but it's just reversed in the plots.

And this is the case also in the main paper.

Am I right?

Hello,

I am looking at the analysis you did for the PBMC with their matched counterparts from GEO (inside the PBMC-prepro.ipynb).

The files used to load the micro_array data seem to be pre processed from the original geo links and was wondering if you can share them within the repository ?

Thank you!
Jay