machenslab
Explained variance calculations
Been using dPCA for a little bit and noticed when I compared to PCA(whether doing a supervised group average PCA or a trial average PCA), that the cumulative variance explained aren't comparable when data is normalized.
Looking further I noticed the issue that the explained variance ratio calculation in python is incorrect, also outlined in a separate issue here https://github.com/machenslab/dPCA/issues/32 At the very least, it doesn't match the matlab calculation, so at least one of these(python or matlab implementation) is wrong.
Lastly, when I run significance testing, the explained variance ratios are updated, even though I don't think this is the desired behavior. Not necessarily an issue, but something to be aware of if anyone else is considering using this method in place of PCA.
