Search All Patents in AI Enabled Search
Patent US10277715
Issued 2019-04-30
Delivering Content Items Using Machine Learning Based Prediction Of User Actions
An online system trains a machine learning model for providing content items to users of the online system. The online system logs actions performed by users and generates user feature vectors based on the logged actions. The online system generates item feature vectors based on information about items from third parties. The machine learning model is trained using the user feature vectors and item feature vectors. The machine learning model determines a likelihood that a target user will acquire a certain item, especially after an update or change has occurred relating to that item. The online system selects content items that the target user is likely to be interested in and is likely to interact with.
Much More than Average Length Specification
View the Patent Matrix® Diagram to Explore the Claim Relationships
USPTO Full Text Publication >
- 1. A method comprising:
receiving, by an online system associated with a plurality of third party systems, information describing a search query specifying a search criteria, wherein the search criteria comprises an item type, wherein the search query requests items of the item type; determining a match score indicative of the degree of match between the search criteria specified by the search query and items returned as search results in response to the search query; storing in association with the user, the search query and the match score of the items returned as search results in response to the search query; for each of one or more third party systems:
receiving information describing one or more items of the item type,
determining a match score for each of the one or more items of the item type and the stored search query, and
determining whether any of the one or more items has a match score greater than the stored match score; and
responsive to identifying an item from a third party system, the item having a match score with the stored search query that is greater than the stored match score, sending, a content item associated with the identified item to a client device of the user.
- 7. A method comprising:
receiving user information about a target user of an online system, the user information describing interactions performed by the target user on the online system; receiving third party information from a third party system, the third party information describing an item associated with the third party that the target user is interested in acquiring; retrieving, a trained machine learning model configured to receive as input the user information and the third party information describing the item and determine a likelihood that a user will perform an action associated with the item; receiving item information indicating a change in a status of the item; determining, using the trained machine learning model and the item information, a likelihood that the target user will perform an action associated with the item; and providing, in response to the likelihood that the target user will perform the action associated with the item exceeding a threshold value, a content item associated with the item for display to the target user.
- 14. A non-transitory computer readable medium storing executable computer program instructions, the computer program instructions comprising instructions that when executed cause a computer processor to:
receive, by an online system associated with a plurality of third party systems, information describing a search query specifying a search criteria, wherein the search criteria comprises an item type, wherein the search query requests items of the item type; determine a match score indicative of the degree of match between the search criteria specified by the search query and items returned as search results in response to the search query; store in association with the user, the search query and the match score of the items returned as search results in response to the search query; for each of one or more third party systems:
receive information describing one or more items of the item type,
determine a match score for each of the one or more items of the item type and the stored search query, and
determine whether any of the one or more items has a match score greater than the stored match score; and
responsive to identifying an item from a third party system, the item having a match score with the stored search query that is greater than the stored match score, send, a content item associated with the identified item to a client device of the user.
