# GroupLens
**Entity class:** Collaborative filtering research system
## Definition and history
GroupLens brought collaborative filtering to the reading of Usenet news through ratings supplied by other participants. [[articles/Intelligence Inside|Intelligence Inside]] situates it after Tapestry in the early 1990s selection lineage. Its purpose was to predict which articles might interest a reader amid a large stream of available material. The signal came from participants' evaluations rather than the article's text alone.
## Mechanism and limits
A collaborative system learns from overlap among people's evaluations, allowing one person's reaction to alter another person's likely encounter. Those evaluations can make discovery more efficient while also creating a feedback loop between exposure and subsequent ratings. Tapestry's explicitly writable filter and GroupLens's rating-based prediction are related but different designs.
## Relationships
**Precursor:** [[wiki/Tapestry|Tapestry]]. **Method:** [[wiki/Collaborative Filtering|Collaborative Filtering]]. **Parallel successor:** [[wiki/Ringo Recommender|Ringo Recommender]]. **Article route:** [[articles/Intelligence Inside|Intelligence Inside]].
## Sources
[ACM, GroupLens paper](https://doi.org/10.1145/192844.192905); [[articles/Intelligence Inside|Intelligence Inside]].