How Not To Get Fooled (even by yourself)


Fitting data is another inductive problem. Here is some data - if we want to control the response (Y-axis) we need to figure out what the relationship between the control (X-axis) and the response is (at least here I'm assuming we've got a pretty good idea which variable is control and which is response - the control is the variable you can change by adjusting a knob or something like that).

We want to come up with a theory that allows us to get the response we want using the control. But how do we do that, and how do we make sure we've done it right?

 Previous Slide Next Slide