Plane Reading: Intro to Statistical Learning

I’ve been working through this book for a while and wanted to post my notes to date. Overall, this book is pretty good — there are areas in which I wish they’d be a little more explicit about the proofs and math. I think they do a good job giving an overview and laying out some of the pros and cons of each model.

The photos are purple because I was on a Virgin America Flight :)

  • Interpretability vs. Accuracy: It seems like in modeling for more business analytics, picking interpretability would be more important. We want to know the levers you can pull. I also wonder about using a model for accuracy if you have a serious consequence when wrong — for example, if you are predicting the patient won’t have a serious adverse reaction to a medication but they will.
  • kNN Regression: There is such a thing as KNN regression — it sounds like you predict what will happen based off “near” points rather than all points, which sounds like an interesting concept
  • Linear Regression with Indicators: is a thing you can do
  • History: The history of statistical methods I thought was interesting

Curious — why talk in odds? It’s a bit confusing to me to do so.

Show your support

Clapping shows how much you appreciated Erica Dohring’s story.