Calling Bullshit: The Art of Skepticism in a Data-Driven World

by Carl T. Bergstrom

If you have ten thousand genes you want to incorporate into your model, good luck in finding the millions of example patients that you will need to have any chance of making reliable predictions.