Clarify scaling description and pipelines
PCA is NOT scale invariant. https://alliance.seas.upenn.edu/~cis520/dynamic/2020/wiki/index.php?n=Lectures.ScaleInvariance https://scikit-learn.org/stable/auto_examples/preprocessing/plot_scaling_importance.html
-
Ensure correct scaling in PCA pipelines (should be already there), logistic regression, SVC or KNeighborsClassifier classification, Ridge/Lasso/ElasticNet regression, or any classification or regression w/ a regularization penalty; -
extend comment in scaling on when to scale (SVMs, kNNs but also PCA or any regularized regression, logistic regression)
Edited by Ghost User