Recommender system
AssigneeMicrosoft Technology Licensing, LLC
InventorsOren Barkan, Noam Koenigstein, Eylon Yogev, Nir Nice
Life of this patent
Estimating the term…
Abstract
A recommender system that represents items in a catalog by first feature vectors in a first vector space based on first characteristics of the items and second feature vectors in a second vector space based on second characteristics of the items different from the first characteristics and maps a feature vector defined in the first vector space for an item to a vector in the second vector space to provide recommendations based on the item.
Claims20
A method of recommending items to a user, the method comprising:
- representing an item by a first feature vector in a first vector space based on first characteristics of the item;
- mapping the first vector to a second vector in a second vector space in which feature vectors representing items are based on second characteristics different from the first characteristics;
- and
- using a relationship between the second vector to another second vector in the second vector space representing another item to determine whether or not to recommend the other item to the user.
A recommender system for recommending an item to a user, the recommender system comprising:
- a database comprising a plurality of catalog items characterized first characteristics and different second characteristics and having second feature vectors in a second vector space that represent the catalog items based on the respective second characteristics of the catalog items;
- and
- a processor configured to:
- receive first characteristics for an item;
- generate a first feature vector in a first vector space different from the second vector space based on the received first characteristics;
- map the generated first feature vector to a second feature vector for the item in the second vector space;
- and
- use the mapped second feature vector and second feature vectors representing items in the catalog to recommend an item from the catalog to the user.
Dependencies are inferred from the claim text.