TEMPORAL CLUSTERING OF SOCIAL NETWORKING CONTENT

• US 20170270180A1
• Filed: 03/15/2016
• Published: 09/21/2017
• Est. Priority Date: 03/15/2016
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: State, Bogdan
• Granted Patent: US 10,795,916 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06F 16/9537 Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
• CPC Class Codes: G06F 40/169 Annotation, e.g. comment data or footnotes
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

The present disclosure is directed towards systems and methods for efficiently identifying periodic trends of user-defined annotations among users of a social networking system. For example, systems and methods described herein involve grouping a set of digital content items into subsets of digital content items and performing one or more iterations of a clustering algorithm on the grouped set of digital content items to identify one or more temporal clusters of the user-defined annotation. Additionally, the systems and methods described herein involve performing one or more additional iterations of the clustering algorithm on one or more reduced sets of digital content items to identify one or more additional temporal clusters of the user-defined annotation. Further, the systems and methods involve determining that the user-defined annotation is periodic based on a time period between identified temporal clusters.

Claims

1. A method comprising:

grouping a set of digital content items associated with timestamps and having a user-defined annotation into randomized subsets of digital content items having the user-defined annotation;

performing, by at least one processor and for each of the randomized subsets of digital content items, a first iteration of a clustering algorithm to identify one or more temporal clusters of the user-defined annotation for the set of digital content items;

grouping a reduced set of digital content items associated with timestamps having the user-defined annotation into reduced subsets of digital content items having the user-defined annotation;

performing, by the at least one processor and for each of the reduced subsets of digital content items, a second iteration of the clustering algorithm to identify one or more additional temporal clusters of the user defined annotation for the reduced set of digital content items; and

determining that the user-defined annotation is periodic based on a time period between the one or more temporal clusters and the one or more additional temporal clusters.

2. The method as recited in claim 1, wherein grouping the set of digital content items into randomized subsets of digital content items comprises generating groups of randomized instances of the user-defined annotation, the randomized instances comprising random timestamps associated with the set of digital content items over a defined period of time.

3. The method as recited in claim 1, wherein the set of digital content items comprises a collection of posts shared with one or more users of a social networking system, the collection of posts comprising timestamps that span over a defined period of time.

4. The method as recited in claim 3, wherein the user-defined annotation comprises a hashtag associated with each post within the collection of posts.

5. The method as recited in claim 1, wherein performing the clustering algorithm to identify one or more temporal clusters within the set of digital content items comprises performing a density-based spatial clustering of applications with noise algorithm on each of the randomized subsets of digital content items to identify one or more temporal clusters for each of the randomized subsets of digital content items.

6. The method as recited in claim 1, further comprising generating the reduced set of digital content items, wherein generating the reduced set of digital content items comprises removing a plurality of instances of the user-defined annotation from the set of digital content items, the plurality of instances corresponding to instances of the user-defined annotation within the identified one or more temporal clusters.

7. The method as recited in claim 6, wherein removing the plurality of instances of the user-defined annotation from the set of digital content items comprises removing each instance of the set of digital content items from the set of digital content items having a timestamp that falls within a range of timestamps corresponding to the identified one or more temporal clusters.

8. The method as recited in claim 6, wherein removing the plurality of instances of the user-defined annotation from the set of digital content items comprises removing all instances of the set of digital content items having a later timestamp than a cutoff point prior to a first identified temporal cluster of the identified one or more temporal clusters.

9. The method as recited in claim 1, wherein:

performing the first iteration of the clustering algorithm further comprises classifying a group of digital content items outside a range of the one or more temporal clusters as noise; and

performing the second iteration of the clustering algorithm comprises performing the clustering algorithm on the group of digital content items identified as noise without considering any instances of the set of digital content items within the range of the one or more temporal clusters.

10. The method as recited in claim 1, wherein determining that the user-defined annotation is periodic comprises:

calculating time periods between each of the one or more temporal clusters and the one or more additional temporal clusters;

comparing the calculated time periods to one or more known time periods; and

identifying a period-type based on comparing the time periods to the one or more known time periods.

11. The method as recited in claim 10, wherein the period-type comprises one of a daily, weekly, monthly, or yearly interval.

12. The method as recited in claim 10, wherein determining that the user-defined annotation is periodic further comprises determining that a number of the one or more temporal clusters and the one or more additional temporal clusters exceeds a threshold number of clusters specific to the identified period-type.

13. The method as recited in claim 1, further comprising adding the user-defined annotation to a blacklist of annotations based on determining that the user-defined annotation is periodic.

14. The method as recited in claim 13, further comprising:

identifying one or more additional user-defined characteristics related to the user-defined characteristic; and

adding the one or more additional user-defined characteristics to the blacklist of characteristics.

15. A system comprising:

at least one processor; and

at least one non-transitory computer readable storage medium storing instructions thereon, that, when executed by the at least one processor, cause the system to;

group a set of digital content items associated with timestamps and having a user-defined annotation into randomized subsets of digital content items having the user-defined annotation;

perform, for each of the randomized subsets of digital content items, a first iteration of a clustering algorithm to identify one or more temporal clusters of the user-defined annotation for the set of digital content items;

group a reduced set of digital content items associated with timestamps having the user-defined annotation into reduced subsets of digital content items having the user-defined annotation;

perform, for each of the reduced subsets of digital content items, a second iteration of the clustering algorithm to identify one or more additional temporal clusters of the user defined annotation for the reduced set of digital content items; and

determine that the user-defined annotation is periodic based on a time period between the one or more temporal clusters and the one or more additional temporal clusters.

16. The system of claim 15, wherein grouping the set of digital content items into randomized subsets of digital content items comprises generating groups of randomized instances of the user-defined annotation, the randomized instances comprising random timestamps associated with the set of digital content items over a defined period of time.

17. The system of claim 15, wherein:

the set of digital content items comprises a collection of posts shared with one or more users of a social networking system, the collection of posts comprising timestamps that span over a defined period of time; and

the user-defined annotation comprises a hashtag associated with each post within the collection of posts.

18. A non-transitory computer readable medium storing instructions thereon that, when executed by at least one processor, cause a computer system to:

group a set of digital content items associated with timestamps and having a user-defined annotation into randomized subsets of digital content items having the user-defined annotation;

perform, for each of the randomized subsets of digital content items, a first iteration of a clustering algorithm to identify one or more temporal clusters of the user-defined annotation for the set of digital content items;

group a reduced set of digital content items associated with timestamps having the user-defined annotation into reduced subsets of digital content items having the user-defined annotation;

perform, for each of the reduced subsets of digital content items, a second iteration of the clustering algorithm to identify one or more additional temporal clusters of the user defined annotation for the reduced set of digital content items; and

determine that the user-defined annotation is periodic based on a time period between the one or more temporal clusters and the one or more additional temporal clusters.

19. The non-transitory computer readable medium of claim 18, wherein grouping the set of digital content items into randomized subsets of digital content items comprises generating groups of randomized instances of the user-defined annotation, the randomized instances comprising random timestamps associated with the set of digital content items over a defined period of time.

20. The non-transitory computer readable medium of claim 18, wherein:

the set of digital content items comprises a collection of posts shared with one or more users of a social networking system, the collection of posts comprising timestamps that span over a defined period of time; and

the user-defined annotation comprises a hashtag associated with each post within the collection of posts.