Open course ware

[AUDIO LOGO] [MOUSE CLICK] STEVEN JOHNSON: OK, so let's get started. So as I said, I'm going to do second derivatives, which is just going to be the derivative of the derivative. When you have functions from matrices to matrices or something like that, you have to think a little bit carefully just to make sure we understand what kind of thing the second derivative is. So remember when we take basically the first derivative, what we have is this linear operator, f primed of x, that takes in a little change and gives us a df, which is the f of x plus dx minus f of x to first order, dropping higher-order terms. And so it's really natural to define the second derivative as the derivative of that. So what should f double prime be? f double primed of x should take in-- let me call it dx prime. I mean, that's not a derivative. It just means a different-- we put it in here in red. This is going to be a different small change. So prime here is not derivative, it...

[CLICK] [SQUEAK] [PAGES RUSTLING] [MOUSE DOUBLE-CLICKS] PROFESSOR: So today we'll be continuing along the theme of risk stratification. I'll spend the first half to 2/3 of today's lecture continuing where we left off last week before the discussion. I'll talk about how does one derive the labels that one uses within a supervised machine learning approach. I'll continue talking about how one evaluates risk stratification models. And then I'll talk about some of the subtleties that arise when you want to use machine learning for health care, specifically for risk stratification. And I think that's going to be one of the most interesting parts of today's lecture. In the last third of today's lecture, I'll be talking about how one can rethink the supervised machine learning problem, not to be a classification problem, but be something closer to a regression problem. And one now thinks about not will someone, for example, develop diabete...