Personalized content delivery system
View Patent ↗A content delivery system for generating personalized content for a user. The system maintains an interest graph that shows the user's current attachment to one or more topics. When a user performs an action, a topic is determined for the action and the user's interest graph is modified based on the action. The system also receives content and analyzes the language of the content to determine a topic of the content. A similarity between the user's interests and the content is determined. The content is also analyzed to determine the popularity of the content. The user's interest level and the popularity of the content are then used to provide the user with a personalized content, such as a content recommendation or enhanced content.
1. A computer implemented method comprising:
receiving information about an action performed by a user, wherein the action comprises at least one of creating or interacting with first content, the user associated with an interest graph stored in a machine readable medium and representative of interests of the user in a plurality of topics;
analyzing language of the first content to identify lexical features of the first content;
determining a topic of the first content based on the lexical features of the first content; and
updating the interest graph stored in the machine readable medium to modify an attachment level in the interest graph representing an interest of the user in the topic of the first content.
2. The method of claim 1 , wherein determining a topic of the first content based on the lexical features of the first content comprises:
comparing the lexical features of the first content to stored lexical features,
wherein the stored lexical features are associated with one or more topics; and
determining the topic of the first content based on the comparison.
3. The method of claim 1 , wherein determining a topic of the first content based on the lexical features of the first content further comprises:
determining a topic of the first content based on the lexical features of the first content and the interest graph associated with the user.
4. The method of claim 1 , wherein determining a topic of the first content based on the lexical features of the first content comprises:
determining a confidence score for the topic of the first content, and
wherein updating the interest graph comprises modifying the attachment level by an amount that is based on the confidence score.
5. The method of claim 1 , wherein updating the interest graph comprises modifying the attachment level based on a type of the received action.
6. The method of claim 1 , further comprising
extracting the first content into a normalized format and
wherein analyzing language of the first content to identify lexical features of the first content comprises analyzing language of the normalized format of the first content.
7. The method of claim 1 , further comprising
generating content for the user based on the interest graph associated with the user.
8. The method of claim 7 , wherein generating content for the user comprises:
generating a content recommendation for the user based on the interest graph associated with the user.
9. The method of claim 7 , further comprising
obtaining second content;
analyzing language of the second content to identify lexical features of the second content;
determining one or more topics of the second content based on the lexical features of the second content; and
determining a similarity between the interests of the user as indicated by the interest graph and the one or more topics of the second content;
wherein the content for the user is generated based on the determined similarity between the interests of the user and the one or more topics of the second content.
10. The method of claim 9 , wherein determining one or more topics of the second content comprises:
comparing the lexical features of the second content to stored lexical features, wherein each of the stored lexical features is associated with one or more topics; and
determining one or more topics of the second content based on the comparison.
11. The method of claim 9 , further comprising:
determining a popularity of the second content, and
wherein the content for the user is generated based on the determined popularity of the second content and the determined similarity between the interests of the user and the one or more topics of the second content.
12. The method of claim 9 , further comprising:
predicting future interests of the user based on the interests of the user as indicated by the interest graph and historical interests of other users, and
wherein determining a similarity comprises determining a similarity between the predicted future interests of the user and the one or more topics of the second content.
13. A computer program product comprising a non-transitory computer-readable medium containing computer program code for performing the method comprising:
receiving information about an action performed by a user, wherein the action comprises at least one of creating or interacting with first content, the user associated with an interest graph and representative of interests of the user in a plurality of topics;
analyzing language of the first content to identify lexical features of the first content;
determining a topic of the first content based on the lexical features of the first content; and
updating the interest graph to modify an attachment level in the interest graph representing an interest of the user in the topic of the first content.
14. The computer program product of claim 13 , wherein determining a topic of the first content based on the lexical features of the first content comprises:
comparing the lexical features of the first content to stored lexical features, wherein the stored lexical features are associated with one or more topics; and
determining the topic of the first content based on the comparison.
15. The computer program product of claim 13 , wherein determining a topic of the first content based on the lexical features of the first content further comprises:
determining a topic of the first content based on the lexical features of the first content and the interest graph associated with the user.
16. The computer program product of claim 13 , further comprising
generating content for the user based on the interest graph associated with the user.
17. The computer program product of claim 16 , further comprising
obtaining second content;
analyzing language of the second content to identify lexical features of the second content;
determining one or more topics of the second content based on the lexical features of the second content; and
determining a similarity between the interests of the user as indicated by the interest graph and the one or more topics of the second content;
wherein the content for the user is generated based on the determined similarity between the interests of the user and the one or more topics of the second content.
18. The computer program product of claim 16 , wherein generating content for the user comprises:
generating a content recommendation for the user based on the interest graph associated with the user.