Posts tagged ‘recommend system’

Collaborative Filtering Resources

maintained by Jun Wang

Generally, collaborative filtering (CF) is any algorithm that filters information for a user based on a collection of user profiles. Users having similar profiles may share similar interests. For a user, information can be filtered in/out regarding to the behaviors of his or her similar users.

Users profiles can be collected either explicitly or implicitly. One can explicitly ask users to rate what they have used/purchased. Such a profile is filled explicitly by the users ratings. An implicit profile is based on passive observation and contains users historic interaction data.

The most common usage of CF is to make recommendation. That’s why collaborative filtering is strongly correlated to recommender system in literature, although CF is only one of the methods for recommender system.

In this page, I collected some useful online materials for collaborative filtering research.


Research Software

Data Sets

Explicit Rating Data Sets:

Implicit Rating Data Sets:

  • Audioscrobblers Music Play-list Data-sets.The Audioscrobbler dataset collects the play-lists of the users in a one-line community ( by using a plug-in in the users’ media players such as Winamp, iTunes, XMMS etc. The plug-ins send the title and artist of every song users play to the Audioscrobbler server, which updates the user’s musical profile with the new songs. In the database, the user’s profile is recorded as a form of co-occurrence pair like {userID,itemID} pair. The pair means a user {userID} has played a\ song {itemID}. The dataset can be obtained at
  • AOL Web search query:

Collaborative Filtering Bibliography

1. Pure Collaborative Filtering


Relevance Models

Latent Class Models

Matrix Factorization


Transitive Associations

Trust Inference


  • Online ranking/collaborative filtering using the perception algorithm (2003).

2. Combining Content-based and Collaborative Filtering

3. Distributed Collaborative Filtering

4. Other issues

Related Information Retrieval Papers

In general, collaborative filtering is formulated as a self-contained problem, apart from classic approaches for text retrieval, e.g. RSJ models and language models. However, the collaborative filtering problem can be treated as a prediction problem – a prediction of the relevance between user and item (see user-item relevance models). Under this veiw, the instant benefits are gained from the current advances in these text retrieval models. We found the following papers are pretty interesting and are related to the collaborative filtering problem.

Related Machine Learning Papers

zz from: