Topic based user profiles
A system and method for developing a user's profile based on the user's interaction with content items. A module on the client rendering the content items or the service including the content items tracks the user's interactions with the content items and transmits the tracked data to a user analysis module. The user analysis module determines the topics associated with the interacted upon content items. The user analysis module then selects the topics for the user's profiles based on the received tracked data and the associated topics. The selected topics are stored in association with the user profile.
1. A computer-implemented method for developing a user profile of a target user, the method comprising:
determining associated topics for each of a plurality of digital content items;
retrieving access data indicating interactions of the target user with at least one of the digital content items;
selecting profile topics for the target user based on the topics associated with the digital content items and the retrieved access data;
storing the selected profile topics in association with the user profile of the target user;
identifying, from a collection of other user profiles, candidate topics co-occurring with a target profile topic from the user profile of the target user;
calculating, for each candidate topic, a count at which the candidate topic co-occurs with the target profile topic across the collection of user profiles;
ranking, by a computer processor, each candidate topic based on the count at which the candidate topic co-occurs with the target profile topic; and
selecting an additional topic to include in the user profile of the target user from the candidate topics based on the ranked candidate topics.
2. The computer-implemented method of claim 1 , wherein the ranking each candidate topic comprises calculating, for each candidate topic, a second count at which the candidate topic occurs in the collection of other user profiles, the count at which the candidate topic co-occurs with the target profile topic across the collection of user profiles normalized based on the second count.
3. The computer-implemented method of claim 1 , wherein the determined associated topics each has a topic strength indicating a degree of association of the topic with the digital content item, and the method further comprising:
selecting the profile topics based on topic strengths.
4. The computer-implemented method of claim 1 , wherein the determined associated topics each has a usefulness weight indicating how useful a topic is in representing its association with the digital content item, and the method further comprising:
selecting the profile topics based on usefulness weights.
5. The computer-implemented method of claim 4 , wherein the usefulness weight of a topic is based on whether digital content items associated with the topic have objectionable content.
6. The computer-implemented method of claim 4 , wherein the usefulness weight of a topic is based on a frequency with which the topic appears in a video corpus.
7. The computer-implemented method of claim 1 , wherein each interaction of the user with the at least one of the digital content items has an associated interaction strength, and the method further comprising:
selecting the profile topics for the target user based on interaction strengths associated with said digital content item.
8. The computer-implemented method of claim 7 , wherein the interaction strength of at least one user interaction is based on a frequency of the at least one user interaction.
9. The computer-implemented method of claim 7 , wherein the interaction strength of at least one user interaction is based on a duration of the at least one user interaction.
10. The computer-implemented method of claim 7 , wherein the interaction strength of at least one user interaction is reduced based on an amount of time elapsed since the at least one user interaction occurred.
11. The computer implemented method of claim 3 , wherein the ranking of each candidate topic co-occurring with the target profile topic is based in part on the degree of association of the candidate topic with the content item.
12. A computer system for developing a user profile of a target user, the system comprising a non-transitory computer readable medium storing instructions for:
determining associated topics for each of a plurality of digital content items;
retrieving access data indicating interactions of the target user with at least one of the digital content items;
selecting profile topics for the target user based on the topics associated with the digital content items and the retrieved access data;
storing the selected profile topics in association with the user profile of the target user;
identifying, from a collection of other user profiles, candidate topics co-occurring with a target profile topic from the user profile of the target user;
calculating, for each candidate topic, a count at which the candidate topic co-occurs with the target profile topic across the collection of user profiles;
ranking each candidate topic based on the count at which the candidate topic co-occurs with the target profile topic; and
selecting an additional topic to include in the user profile of the target user from the candidate topics based on the ranked candidate topics.
13. The computer system of claim 12 , wherein the ranking each candidate topic comprises calculating, for each candidate topic, a second count at which the candidate topic occurs in the collection of other user profiles, the count at which the candidate topic co-occurs with the target profile topic across the collection of user profiles normalized based on the second count.
14. The computer system of claim 12 , wherein the determined associated topics each has a topic strength indicating a degree of association of the topic with the digital content item, and the instructions further comprising:
selecting the profile topics are further selected based on topic strengths.
15. The computer system of claim 12 , wherein the determined associated topics each has a usefulness weight indicating how useful a topic is in representing its association with the digital content item, and the instructions further comprising:
selecting the profile topics based on usefulness weights.
16. The computer system of claim 15 , wherein the usefulness weight of a topic is based on whether digital content items associated with the topic have objectionable content.
17. The computer system of claim 15 , wherein the usefulness weight of a topic is based on a frequency with which the topic appears in a video corpus.
18. The computer system of claim 12 , wherein each interaction of the user with the at least one of the digital content items has an associated interaction strength, and the instructions further comprising:
selecting the profile topics for the target user based on interaction strengths associated with said digital content item.
19. The computer system of claim 18 , wherein the interaction strength of at least one user interaction is based on a frequency of the at least one user interaction.
20. The computer system of claim 18 , wherein the interaction strength of at least one user interaction is based on a duration of the at least one user interaction.
21. The computer system of claim 18 , wherein the interaction strength of at least one user interaction is reduced based on an amount of time elapsed since the at least one user interaction occurred.
22. The computer system of claim 14 , wherein the ranking of each candidate topic co-occurring with the target profile topic is based in part on the degree of association of the candidate topic with the content item.