US 10,311,362 B1Granted

Identification of trending content using social network activity and user interests

AssigneeAmazon Technologies, Inc.

InventorsMuneer Mirza, Saral Jain, Ameet Nirmal Vaswani

Life of this patent

Estimating the term…

Loading patent…

Abstract

Network accessible content, including but not limited to web pages, can be analyzed to identify characteristics of the content. Historical content requests associated with users and/or other user-specific information can be analyzed to identify user interests. Content can then be provided or recommended to users based on content characteristics and user interests. The content that is provided or recommended may be trending content that has at least a threshold level of popularity, became popular recently, and/or is associated with a threshold level of social network activity.

Claims17

Dependent claims

A system comprising one or more physical computing devices, the system configured to at least:

obtain content request data regarding requests, initiated during a period of time, for a plurality of different content items, the requests generated by user computing devices and containing uniform resource locators (URLs) of requested content items;
modify the content request data at least partly by converting the URLs in at least some of the requests to canonical URLs, wherein converting a URL of a content item to a canonical URL comprises discarding URL text unnecessary for identifying the content item;
identify, using the modified content request data, a set of content items that have been requested at least a threshold number of times during the period of time;
generate scores for content items in the set of content items, wherein a score for a first content item of the set of content items is based at least partly on a difference between a current time and a time at which the first content item first reaches a threshold level of popularity;
select a subset of the content items from the set of content items based at least partly on the scores;
select a set of trending content items from the subset of content items based at least partly on information, obtained from one or more social network servers, reflecting an amount of social network activity associated with individual content items of the set of trending content items in the period of time;
and
generate a list of user-specific trending content items for a user, the list of user-specific content items generated from the set of trending content items based at least partly on an analysis of content interests associated with the user and corresponding characteristics of content items in the list of user-specific trending content items.

5 more elements

A system comprising:

a data repository that stores a record of content requests made by user devices to content servers, the record of content requests including uniform resource locators (URLs) included in the content requests, the URLs specifying requested content items;
and
a computing system comprising one or more computing devices, the computing system programmed to identify trending content items by a process that comprises:
generating a modified record of content requests at least partly by converting URLs in the record of content requests to canonical URLs, wherein converting a URL of a content item to a canonical URL comprises discarding URL text unnecessary for identifying the content item;
determining, based on the modified record of content requests, rates of change in popularity of particular content items;
selecting, based at least partly on the rates of change in popularity, a subset of the content items that satisfy a popularity condition;
retrieving, for individual content items in the subset, social network interaction data indicating levels to which the respective content items have been the subject of social network activity;
and
selecting, from the subset of content items, based at least partly on the social network interaction data, particular content items that are trending.

6 more elements

Dependencies are inferred from the claim text.