Create the free instant account at recombee.com now!
The recommendation domain consists of three principal components:
We assume that there is a catalog of items that are available to users. Users may interact with the items:
Besides their ID, the items and users may have certain properties. The properties are analyzed by the recommender to further improve the recommendations. Also, you may use the properties to filter or boost recommendations of some specific items. The set of properties is managed by the API user, who may define properties according to the target domain. The properties may be thought as columns in relational database table: they are of certain types and they may be of specific values for different items.
Sets of items with natural ordering, such as consecutive book parts or TV show episodes, can be organized into Series.
Recommendation endpoints can return items, which is the most common case, or users.
We provide client libraries (SDKs) for easy use of our API.