> But false positives are false positives; why does it matter how
> many true positives there were? Because it's a measure of how muddy the
> water is? It seems like here, absolute numbers of false positives would
> be more valuable in many situations. As Google found, it often doesn't
> matter how many false positives you have, as long as the most valuable
> true positives are close to the top of the list.
I doubt most people look much beyond the first few pages of two Google, whether the results are good or not, so it would seem reasonable to frame ranking as filtering and call the first few pages the result set. Is it *really* a special case?
> Incidentally, this is not a purely academic line of questioning; I
> worked on an information retrieval project that failed in part because
> precision and recall did not accurately predict customer satisfaction.
Isn't it likely that this was more a problem with whatever proxy you were using for customer satisfaction, than precision and recall per se? If, for example, you were measuring "number of customers satisfied" against "number of customers served", you ought to find that these sorts of metrics do their job quite well... I'd have thought.
Justin Washtell
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.