Determining trending topics in social media

• US 9,830,386 B2
• Filed: 12/30/2014
• Issued: 11/28/2017
• Est. Priority Date: 12/30/2014
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Huang, Fei, Rottmann, Kay, Zhang, Ying, Eck, Matthias Gerhard
• Primary Examiner(s): Vy, Hung T
• Application Number:
  US14/586,049
• Publication Number:
  US 20160188703A1
• Time in Patent Office: 1,064 Days
• Field of Search: 707750, 707772, 707731
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: G06Q 50/10 Services

Abstract

Technology is discussed herein for identifying comparatively trending topics between groups of posts. Groups of posts can be selected based on parameters such as author age, location, gender, etc., or based on information about content items such as when they were posted or what keywords they contain. Topics, as one or more groups of words, can each be given a rank score for each group based on the topic'"'"'s frequency within each group. A difference score for selected topics can be computed based on a difference between the rank score for the selected topic in each of the groups. When the difference score for a selected topic is above a specified threshold, that selected topic can be identified as a comparatively trending topic.

Claims

1. A method, performed by a computing system, for identifying comparatively trending topics, comprising:

receiving, by the server, as post parameters to an interface of the computing system, one or more parameters that distinguish social media postings;

obtaining, by the server, over a network and based on the received post parameters, two groups of posts;

generating, by the server, a first ranked list of topics for a first group of the two groups of posts by;

identifying, by the server, as first group topics, topics from the first group of posts wherein members of the first group topics each comprise one or more words; and

computing, by the server, a first rank score for each of the topics of the first group topics;

generating, by the server, a second ranked list of topics for a second group of the two groups of posts by;

identifying, by the server, as second group topics, topics from the second group of posts wherein members of the second group topics each comprise one or more words; and

computing, by the server, a second rank score for each of the topics of the second group topics;

associating, by the server, with a selected one of the first group topics from the first ranked list of topics, a difference score computed, by the computing system, as a difference between;

the rank score for the selected one of the first group topics, and

a corresponding rank score determined based on the second ranked list of topics;

comparing, by the server, the difference score to a specified threshold;

in response to determining that the difference score is above the specified threshold, identifying, by the server, the selected one of the first group topics as a comparatively trending topic.

2. The method of claim 1 wherein the post parameters comprise at least one of:

one or more keywords;

one or more topics;

post length or a post length range;

one or more languages;

one or more specific authors;

one or more times or time frames;

author age or author age range;

one or more author genders;

one or more author locations;

one or more author biometrics;

one or more identified friend networks which an author is a part of;

a history of one or more types of content items which authors historically interact with;

one or more source devices;

or

one or more IP address or ranges of IP addresses.

3. The method of claim 1 wherein obtaining, based on the received post parameters, two groups of posts is performed by:

selecting the first group using the received one or more parameters; and

selecting the second group as a random selection of posts.

4. The method of claim 1 wherein receiving post parameters comprises:

receiving a first subset of the post parameters with corresponding parameter types;

receiving an alternate parameter with a specified type matching one of the corresponding parameter types; and

creating a second subset of the post parameters with of the parameters of the first subset except for the parameter of the first subset with the type matching the specified type replaced by the alternate parameter; and

wherein obtaining, based on the received post parameters, two groups of posts is performed by;

selecting the first group using the first subset of the post parameters; and

selecting the second group using the second subset of the post parameters.

5. The method of claim 1 wherein the first group topics and the second group topics are each an n-gram of no more than a specified number of words.

6. The method of claim 1 wherein a rank score, of the first rank scores, for a selected topic is computed by:

computing a term frequency of the selected topic within the first group;

computing an inverse document frequency of the selected topic within the first group; and

computing the rank score for the selected topic as a product of the term frequency and inverse document frequency.

7. The method of claim 1 wherein the corresponding rank score is determined by:

locating a matching topic from the second group topics, the matching topic corresponding to the selected one of the first group topics;

wherein the corresponding rank score is the second rank score which was computed for the matching topic from the second group topics.

8. The method of claim 1 wherein the corresponding rank score is determined by:

determining that the second group topics do not include a topic corresponding to the selected one of the first group topics; and

in response to determining that the second group topics do not include the topic corresponding to the selected one of the first group topics, using a default value as the corresponding rank score.

9. The method of claim 1 further comprising:

receiving an identification of a common language; and

translating any of the first group topics and any of the second group topics which are not in the common language to be in the common language.

10. The method of claim 1 further comprising:

generating a common word list comprising topics that have the first rank score above a specified common threshold and the second rank score above the specified common threshold.

11. The method of claim 1 further comprising:

receiving a list of common topics determined to be too common to be useful as comparatively trending topics;

removing, from the first ranked list of topics, topics appearing on the list of common topics; and

removing, from the second ranked list of topics, topics appearing on the list of common topics.

12. A non-transitory computer-readable storage medium storing instructions configured to, when executed by a computing system, cause the computing system to perform operations for identifying comparatively trending topics, the operations comprising:

receiving, as post parameters, one or more parameters that distinguish social media postings;

obtaining, over a network and based on the received post parameters, at least two groups of posts;

generating, for each selected group of posts of the at least two groups of posts, a corresponding ranked list of topics by;

identifying topics from the selected group of posts wherein each identified topic comprises one or more words; and

computing a rank score for each of the identified topics;

associating, with an individual topic from a selected one of the corresponding ranked lists, a difference score computed as a difference between;

the rank score for the individual topic from the selected one of the corresponding ranked lists, and

a rank score for the individual topic computed from one or more of the ranked lists other than the selected one of the corresponding ranked lists;

comparing the difference score to a specified threshold;

in response to determining that the difference score is above the specified threshold, identifying the individual topic as a comparatively trending topic.

13. The non-transitory computer-readable storage medium of claim 12 wherein the post parameters comprise at least one of:

one or more keywords;

one or more topics;

one or more times or time frames;

author age or author age range;

one or more author genders;

or

one or more author locations.

14. The non-transitory computer-readable storage medium of claim 12 wherein receiving post parameters comprises:

receiving a first subset of the post parameters with corresponding parameter types;

receiving an alternate parameter with a specified type matching one of the corresponding parameter types; and

creating a second subset of the post parameters with of the parameters of the first subset except for the parameter of the first subset with the type matching the specified type replaced by the alternate parameter; and

wherein obtaining, based on the received post parameters, at least two groups of posts is performed by;

selecting a first of the at least two groups of posts using the first subset of the post parameters; and

selecting a second of the at least two groups of posts using the second subset of the post parameters.

15. The non-transitory computer-readable storage medium of claim 12 wherein a selected rank score for a selected topic, of the rank scores for each of the identified topics, is computed by:

computing a term frequency of the selected topic within a first of the at least two groups;

computing an inverse document frequency of the selected topic within the first of the at least two groups; and

computing the selected rank score as a product of the term frequency and inverse document frequency.

16. The non-transitory computer-readable storage medium of claim 12, wherein the operations further comprise:

receiving an identification of a common language; and

translating any of the posts of the least two groups of posts which are not in the common language to be in the common language.

17. The non-transitory computer-readable storage medium of claim 12, wherein the operations further comprise:

receiving a list of common topics determined to be too common to be useful as comparatively trending topics; and

removing, from each ranked list of topics, topics appearing on the list of common topics.

18. A computing system for identifying comparatively trending topics, comprising:

one or more processors;

a memory;

an input configured to receive, as post parameters, one or more parameters that distinguish social media postings;

a transceiver configured to obtain, based on the received post parameters, two groups of posts;

a group builder configured to;

generate a first ranked list of topics for a first group of the two groups of posts by;

identifying, as first group topics, topics from the first group of posts wherein members of the first group topics each comprise one or more words; and

computing a first rank score for each of the topics of the first group topics; and

generate a second ranked list of topics for a second group of the two groups of posts by;

identifying, as second group topics, topics from the second group of posts wherein members of the second group topics each comprise one or more words; and

computing a second rank score for each of the topics of the second group topics;

a ranked topic list creator configured to associate, with a selected one of the first group topics, a difference score computed as a difference between;

the rank score for the selected one of the first group topics from the first ranked list of topics, and

a corresponding rank score determined based on the second ranked list of topics; and

a comparatively trending topic identifier configured to;

compare the difference score to a specified threshold; and

in response to determining that the difference score is above the specified threshold, identify the selected one of the first group topics as a comparatively trending topic.

19. The system of claim 18 wherein the corresponding rank score is computed by:

locating a matching topic from the second group topics, the matching topic corresponding to the selected one of the first group topics; and

using as the corresponding rank score the second rank score which was computed for the matching topic from the second group topics.

20. The system of claim 18 wherein the input receives the post parameters by:

receiving a first subset of the post parameters with corresponding parameter types;

receiving an alternate parameter with a specified type matching one of the corresponding parameter types; and

creating a second subset of the post parameters with of the parameters of the first subset except for the parameter of the first subset with the type matching the specified type replaced by the alternate parameter; and

wherein the two groups of posts are selected by;

selecting the first group using the first subset of the post parameters; and

selecting the second group using the second subset of the post parameters.