>
http://www.librarything.com/> ... * If the buzz page doesn't convince you,
> you cannot be convinced. Go away.
Since this is a classic recommendation system that
evaluates taste similarity between people based on
what they have bought, if someone knows about a
statistical study that quantifies how relevant
some system(s) like this actually is/are
(based on feedback from users on recommendations
given to them, and the number of books/products they
themselves had to valuate), I'd very much like to
have a reference. More precisely, since there surely
are dozens of such analysis, I'd like someone who
knows about them to refer me to some synthesis or
conclusion.
I am asking because I very rarely got a relevant
automatic recommendation. Is this common/uncommon?
Thanks,
Philippe
---------------------------------------------------------------------
To unsubscribe, e-mail:
cg-unsubscribe@...
For additional commands, e-mail:
cg-help@...
