Collaborative ltering attempts to alleviate in- formation overload by o ering recommenda- tions on whether information is valuable based on the opinions of those who have already eval- uated it. Usenet news is an information source whose value is being severely diminished by the volume of low-quality and uninteresting infor- mation posted in its newsgroups. The Grou- pLens system applies collaborative ltering to Usenet news to demonstrate how we can re- store the value of Usenet news by sharing our judgements of articles, with our identities pro- tected by pseudonyms. This paper extends the original GroupLens work by reporting on a signi cantly enhanced system and the results of a seven week trial with 250 users and over 20,000 news articles. GroupLens has an open and exible architec- ture that allows easy integration of new news- reader clients and ratings bureaus. We show ratings and prediction pro les for three news- groups, and assess the accuracy of the predictions.
|Original language||English (US)|
|State||Published - Jan 1 1997|
|Event||USENIX 1997 Annual Technical Conference - Anaheim, United States|
Duration: Jan 6 1997 → Jan 10 1997
|Conference||USENIX 1997 Annual Technical Conference|
|Period||1/6/97 → 1/10/97|