1 comment

0

I can say that this is a sort of a special case need (in that I've gone years in the industry without needing to do a random sample of a dataset of unknown size; I don't do a lot of statistical sampling; I've cared more about data in aggregate) but I really like this algorithm. It's explained very well.