On: Linear compression in Python: PCA vs unsupervised feature selection

[closed]

That's fair and was helpful to hear! I had dimensionality reduction in the back of my mind, and now that you mention it, your point about conceptual sparsity definitely seems important here.