
Power K-Means
Data Skeptic
00:00
Are There Limits to Inference?
If the data aren't separable in a particular way, then there will be limits to how well you can cluster. No matter how good your optimization problem is, no matter if you are able to show p equals n p, there's still an amount of error that's intrinsic to the statistical noise or the sort of data generating mechanism. So there are absolutely some fundamental limits to inference as well.
Transcript
Play full episode