thanks! i believe its this one? rob.schapire.net/papers/Schapire99... certainly remains a fascinating idea.
Yes, this one, I liked it a lot although it explains the "sampling of the dataset" poorly, in my opinion. In that paper he shows some results on a "letter dataset". Never found it on the internet. If one finds the dataset I would like to have a link to it :)
We're a place where coders share, stay up-to-date and grow their careers.
We strive for transparency and don't collect excess data.