[Home]CollaborativeFiltering

MeatballWiki | RecentChanges | Random Page | Indices | Categories

The use of ratings supplied by multiple users of a system to indicate relative merits of material. While simple systems merely identify a single concensus value, more advanced ones use tools to identify groups of raters with similar (or at least predictive) interests, e.g.: BayesianMethods?.

Collaborative filtering is an example of IncidentalCollaboration. The collaborators each work independently, and sometimes without knowledge of their fellows' ratings.

A more detailed definition is provided by Doug Oard in "Information Filtering Defined" [1]:

Generally, the goal of an information filtering system is to sort through large volumes of dynamically generated information and present to the user those which are likely to satisfy his or her information requirement.

Useful links (and collections of same):

The topic is, for cyberspace, quite old, with roots in the late 1960s. Typical focuses have been Usenet, email filtering, and more recently, webpage filtering. Collaborative moderation systems such as ScoopEngine (used at KuroShin) and SlashCode (SlashDot).

Another feature which may be added someday to systems with moderation is collaborative filtering in the Amazon sense. That is, you could request the system to show you articles that others with similar preferences to you voted for. "Similar preferences" would be extracted from comparing your past voting pattern to others'.

This is strongly related to ContentFiltering.


CategoryCollaboration

Discussion

MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions
Search: