System and method for estimating user ratings from user behavior and providing recommendations
First Claim
1. A recommendation system for providing a recommendation set of recommended items from an item data store to a user who uses an interface to browse an item set made from items in the item data store, the recommendation system comprising:
- a) a ratings estimation module that receives item property data including item attribute data for items in the item set, the ratings estimation module being connected to the interface for receiving training set data including user-specific information and being adapted to generate a training set based on at least one item browsed by the user and, based on at least one user-specific property and one or more item attributes, the ratings estimation module generates an item rating matrix by estimating user ratings for the items in the training set; and
, b) a recommendation module connected to the ratings estimation module for receiving the item rating matrix and the item set, wherein the recommendation module estimates user ratings for unseen items in the item set and generates the recommendation set based on the estimated ratings.
2 Assignments
0 Petitions
Accused Products
Abstract
A recommendation system and method for generating a set of recommended items for a user who browses items from an item data store. The recommendation system estimates ratings for at least some of the items previously viewed by the user, and then constructs a representation of the user'"'"'s implicit ratings of attributes of these items. These implicit attribute ratings are then used to estimate the user'"'"'s ratings for unseen items in the item data store.
-
Citations
36 Claims
-
1. A recommendation system for providing a recommendation set of recommended items from an item data store to a user who uses an interface to browse an item set made from items in the item data store, the recommendation system comprising:
-
a) a ratings estimation module that receives item property data including item attribute data for items in the item set, the ratings estimation module being connected to the interface for receiving training set data including user-specific information and being adapted to generate a training set based on at least one item browsed by the user and, based on at least one user-specific property and one or more item attributes, the ratings estimation module generates an item rating matrix by estimating user ratings for the items in the training set; and
,b) a recommendation module connected to the ratings estimation module for receiving the item rating matrix and the item set, wherein the recommendation module estimates user ratings for unseen items in the item set and generates the recommendation set based on the estimated ratings. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20)
-
-
21. A method for providing a recommendation set of recommended items from an item data store to a user who uses an interface to browse an item set made from items in the item data store, the method including:
-
a) obtaining item property data including item attribute data for items in the item set and training set data including user specific property information;
b) generating a training set based on at least one item browsed by the user and attribute values for the at least one item;
c) generating an item rating matrix by estimating user ratings for items in the training set based on at least one user specific property and one or more item attributes;
d) estimating user ratings for unseen items in the item set based on the item rating matrix; and
,e) generating the recommendation set based on the estimated ratings. - View Dependent Claims (22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36)
-
Specification