If you want to annoy your neighbors, tell the truth about them.
Imagine that you’re trying to predict how I’m going to vote in the next presidential election. If you know nothing else about me (and if you have the data), one sensible approach is to look at how my neighbors are planning to vote. Living in Seattle, as I do, my neighbors are invariably planning to vote for the Democratic candidate, which suggests that “Democratic candidate” is a good guess for me as well.
Now imagine you know more about me than just geography—perhaps you know my age, my income, how many kids I have, and so on. To the extent my behavior is influenced (or characterized) by those things, looking just at my neighbors who are close to me among all those dimensions seems likely to be an even better predictor than looking at all my neighbors. This is the idea behind nearest neighbors classification.
Nearest neighbors is one of the simplest predictive models there is. It makes no mathematical assumptions, and it doesn’t require any sort of heavy machinery. The only things it requires are:
Some notion of distance
An assumption that points that are close to one another are similar
Most of the techniques we’ll see in this book look at the dataset as a whole in order to learn patterns in the data. Nearest neighbors, on the other hand, quite consciously neglects a lot of information, since the prediction for each new point depends only on the handful ...