Intelligent system and methods of recommending media content items based on user preferences
First Claim
Patent Images
1. A distributed system for predicting items for a user, the system comprising:
- a client;
a server in communication with a plurality of clients, including the client, over a network connection;
device logic at the server that periodically receives a list of user-rated items from each client of the plurality of clients, the lists of user-rated items aggregated into a single aggregated list of items, the items being associated with media content, the user-rated items in the list of user-rated items from the client being rated by a user, the user-rated items in the lists of user-rated items from other clients of the plurality of clients being rated by other users;
logic at the server that filters user-rated items from the single aggregated list of items based on frequency by monitoring frequency of the user-rated items and discarding items and the items'"'"' corresponding user-ratings that do not satisfy a threshold frequency;
logic at the server that creates matrices, each matrix of the matrices corresponding to each unique pair of items from the single aggregated list of items, each matrix storing user-ratings for each item of the unique pair of items, the matrices anonymous with respect to the user and the other users;
logic at the server that computes a rating correlation between items of the unique pair of items from each matrix;
logic at the server that filters out non-significant rating correlations of unique pairs of items;
logic at the server that compiles a list of correlating items comprising for which the rating correlations were not filtered out;
logic at the server that periodically sends the list of correlating items to the client; and
logic at the client that predicts a rating for an unrated item based on the correlations provided in the list of correlating items.
12 Assignments
0 Petitions
Accused Products
Abstract
A system and method for making program recommendations to users of a network-based video recording system utilizes expressed preferences as inputs to collaborative filtering and Bayesian predictive algorithms to rate television programs using a graphical rating system. The predictive algorithms are adaptive, improving in accuracy as more programs are rated.
79 Citations
35 Claims
-
1. A distributed system for predicting items for a user, the system comprising:
-
a client; a server in communication with a plurality of clients, including the client, over a network connection; device logic at the server that periodically receives a list of user-rated items from each client of the plurality of clients, the lists of user-rated items aggregated into a single aggregated list of items, the items being associated with media content, the user-rated items in the list of user-rated items from the client being rated by a user, the user-rated items in the lists of user-rated items from other clients of the plurality of clients being rated by other users; logic at the server that filters user-rated items from the single aggregated list of items based on frequency by monitoring frequency of the user-rated items and discarding items and the items'"'"' corresponding user-ratings that do not satisfy a threshold frequency; logic at the server that creates matrices, each matrix of the matrices corresponding to each unique pair of items from the single aggregated list of items, each matrix storing user-ratings for each item of the unique pair of items, the matrices anonymous with respect to the user and the other users; logic at the server that computes a rating correlation between items of the unique pair of items from each matrix; logic at the server that filters out non-significant rating correlations of unique pairs of items; logic at the server that compiles a list of correlating items comprising for which the rating correlations were not filtered out; logic at the server that periodically sends the list of correlating items to the client; and logic at the client that predicts a rating for an unrated item based on the correlations provided in the list of correlating items. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19)
-
-
20. A method of compiling a list of correlated items, the method comprising the steps of:
-
periodically receiving at a server a list of items rated by the user from a client of a plurality of clients over a stateless network connection, the items being associated with media content; aggregating the list of items at the server with lists from other clients of the plurality of clients in contact with the server into a single aggregated list of items, the items in the lists from the other clients rated by other users of a plurality of users; filtering the single aggregated list of items by monitoring frequency of user-rated items, and discarding items and corresponding user-ratings that do not satisfy a threshold frequency; tallying user-ratings for each item of each unique pair of items from the single aggregated list of items and storing tallies in one or more pair matrices, each pair matrix corresponding to each unique pair of items, the one or more pair matrices anonymous with respect to users of the plurality of users; computing a rating correlation between items of the unique pair of items from each pair matrix; filtering out non-significant rating correlations of unique pairs of items; compiling a list of correlating items comprising items for which the rating correlations were not filtered out; and periodically sending by the server the list of correlating items to the client of the plurality of clients. - View Dependent Claims (21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35)
-
Specification