Misc,
Open Rating Systems
(December 2002)
Abstract
In the offline world, we look to the people we trust
and those they trust for reliable information. In this
paper, we present a computational model of this
phenomenon and show how it can be used to identify high
quality content in an Open Rating System, i.e., a
system in which any user can rate content. We present a
case study (Epinions.com) of a system based on this
model and describe a new platform called PeopleNet for
harnessing this phenomenon in an open distributed
fashion.
