While we are at new modules 0:)
After @JoKra1's suggestion I'm working on adding a regularization to the cumulative method using a pseudo-kfold (pseudo in the sense that we do not recompute the PCA). This, along with the LOOCV that you plan @jelmerborst, we could put all of that into a crossvalidation module instead of utils or the evaluation discussed in #279
I could be done by monday for a first draft.
While we are at new modules 0:)
After @JoKra1's suggestion I'm working on adding a regularization to the
cumulativemethod using a pseudo-kfold (pseudo in the sense that we do not recompute the PCA). This, along with the LOOCV that you plan @jelmerborst, we could put all of that into acrossvalidationmodule instead ofutilsor theevaluationdiscussed in #279I could be done by monday for a first draft.