Suggested queries for locating posts on online social networks

• US 10,102,273 B2
• Filed: 12/30/2014
• Issued: 10/16/2018
• 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: Lee-Goldman, Russell, Hazra, Kurchi Subhra
• Primary Examiner(s): Gofman, Alex
• Assistant Examiner(s): Dewan, Kamal K
• Application Number:
  US14/585,782
• Publication Number:
  US 20160188702A1
• Time in Patent Office: 1,386 Days
• Field of Search: 707749, 707708, 707726, 707732, 707755, 707767, 707770, 707990005, 707E17014
• CPC Class Codes: G06F 16/24575 using context
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising

Abstract

In one embodiment, a method includes accessing a post of an online social network, extracting keywords from the content of the first post and the metadata associated with the first post, determining topics associated with each extracted keyword, calculating a topic-score based on a relevance of the topic to the post for each topic, generating a suggested keyword query corresponding to the post, the suggested keyword query comprising extracted keywords corresponding to each topic having a topic-score greater than a threshold topic-score, and sending the post and the suggested keyword query to a client device of a first user for display.

Claims

1. A method for improving search query suggestions comprising, by one or more computing devices:

accessing a first post of an online social network, the first post comprising a content of the first post and a metadata associated with the first post;

extracting a plurality of keywords from the content of the first post and the metadata associated with the first post;

determining, for each extracted keyword of the first post, one or more topics associated with the extracted keyword;

calculating, for each topic associated with the extracted keywords of the first post, a topic-score based on a relevance of the topic to the first post;

determining, from the topics associated with the extracted keywords of the first post, one or more relevant topics for the first post, wherein each relevant topic has a topic-score greater than a threshold topic-score;

selecting a plurality of keywords from the extracted keywords of the first post, each of the selected keywords corresponding to one or more of the relevant topics;

generating a suggested keyword query corresponding to the first post, the suggested keyword query comprising the selected keywords of the first post;

determining that a total number of the selected keywords of the first post in the suggested keyword query is less than a predetermined number of keywords, wherein the predetermined number of keywords is at least two keywords;

in response to determining that the total number of the selected keywords of the first post in the suggested keyword query is less than the predetermined number of keywords, modifying the suggested keyword query to include at least a portion of a name of an author of the first post; and

sending, to a client device of a first user of the online social network, information configured to render an interface comprising the first post and the modified suggested keyword query, wherein the first post is displayed in association with the modified suggested keyword query.

2. The method of claim 1, further comprising:

accessing a social graph comprising a plurality of nodes and a plurality of edges connecting the nodes, each of the edges between two of the nodes representing a single degree of separation between them, the nodes comprising;

a first node corresponding to the first user; and

a plurality of second nodes corresponding to a plurality of posts associated with the online social network, respectively.

3. The method of claim 1, wherein determining, for each extracted keyword, the one or more topics associated with the extracted keyword comprises:

receiving an electronic topic index comprising a set of objects, wherein each object represents a topic;

determining a context of the extracted keyword in the first post;

calculating a context-score for each object of the set of objects based on the determined context; and

determining whether there is an object having a context-score greater than a threshold context-score.

4. The method of claim 3, wherein the context of the extracted keyword in the first post is determined using a forward-backward algorithm to calculate a probability of the extracted keyword matching the topic based on adjacent n-grams that match the topic also being present.

5. The method of claim 1, wherein the relevance of the topic to the first post is based on an overall context of the first post.

6. The method of claim 1, wherein the topic-score is further based on a current popularity of the topic on the online social network.

7. The method of claim 1, wherein the topic-score is further based on a confidence level that a particular extracted keyword associated with the topic corresponds to the topic, the confidence level being based on a context in which the particular extracted keyword is used in the post.

8. The method of claim 1, wherein extracting the plurality of keywords comprises filtering out one or more n-grams from the content of the first post having a low relevance based on a TF-IDF analysis.

9. The method of claim 1, further comprising removing a name of an author of the first post from the extracted keywords.

10. The method of claim 1, wherein the suggested keyword query comprises a fixed number of extracted keywords.

11. The method of claim 1, further comprising sending instructions for display indicating that the modified suggested keyword query corresponding to the first post can be used to find the first post again or to find related posts.

12. The method of claim 1, wherein the modified suggested keyword query is displayed adjacent to the first post.

13. The method of claim 1, wherein the extracted keywords in the modified suggested keyword query are displayed in an order based on a relative position of the extracted keywords in the content of the first post.

14. The method of claim 1, wherein the suggested keyword is a link that is selectable to execute a search query comprising the modified suggested keyword query.

15. The method of claim 1, further comprising verifying that the suggested keyword query retrieves the first post in response to executing the modified suggested keyword query.

16. The method of claim 1, further comprising:

receiving, from the client device of the first user, a search query comprising the modified suggested keyword query; and

generating one or more search results comprising references to one or more posts of the online social network matching the modified suggested keyword query, wherein at least one of the search results comprises a reference to the first post.

17. The method of claim 1, further comprising:

receiving, from the client device of the first user, a search query comprising the modified suggested keyword query; and

generating one or more search results comprising references to one or more second posts of the online social network matching the modified suggested keyword query, wherein at least one of the second posts is a related post.

18. One or more computer-readable non-transitory storage media embodying software for improving search query suggestions, the software operable when executed to:

access a first post of an online social network, the first post comprising a content of the first post and a metadata associated with the first post;

extract a plurality of keywords from the content of the first post and the metadata associated with the first post;

determine, for each extracted keyword of the first post, one or more topics associated with the extracted keyword;

calculate, for each topic associated with the extracted keywords of the first post, a topic-score based on a relevance of the topic to the first post;

determine, from the topics associated with the extracted keywords of the first post, one or more relevant topics for the first post, wherein each relevant topic has a topic-score greater than a threshold topic-score;

selecting a plurality of keywords from the extracted keywords of the first post, each of the selected keywords corresponding to one or more of the relevant topics;

generate a suggested keyword query corresponding to the first post, the suggested keyword query comprising the selected keywords of the first post;

determine that a total number of the selected keywords of the first post in the suggested keyword query is less than a predetermined number of keywords, wherein the predetermined number of keywords is at least two keywords;

in response to determining that the total number of the selected keywords of the first post in the suggested keyword query is less than the predetermined number of keywords, modify the suggested keyword query to include at least a portion of a name of an author of the first post; and

send, to a client device of a first user of the online social network, information configured to render an interface comprising the first post and the modified suggested keyword query, wherein the first post is displayed in association with the modified suggested keyword query.

19. A system comprising:

one or more processors; and

a non-transitory memory coupled to the processors comprising instructions executable by the processors for improving search query suggestions, the processors operable when executing the instructions to;

access a first post of an online social network, the first post comprising a content of the first post and a metadata associated with the first post;

extract a plurality of keywords from the content of the first post and the metadata associated with the first post;

determine, for each extracted keyword of the first post, one or more topics associated with the extracted keyword;

calculate, for each topic associated with the extracted keywords of the first post, a topic-score based on a relevance of the topic to the first post;

determine, from the topics associated with the extracted keywords of the first post, one or more relevant topics for the first post, wherein each relevant topic has a topic-score greater than a threshold topic-score;

selecting a plurality of keywords from the extracted keywords of the first post, each of the selected keywords corresponding to one or more of the relevant topics;

generate a suggested keyword query corresponding to the first post, the suggested keyword query comprising the selected keywords of the first post;

determine that a total number of the selected keywords of the first post in the suggested keyword query is less than a predetermined number of keywords, wherein the predetermined number of keywords is at least two keywords;

in response to determining that the total number of the selected keywords of the first post in the suggested keyword query is less than the predetermined number of keywords, modify the suggested keyword query to include at least a portion of a name of an author of the first post; and

send, to a client device of a first user of the online social network, information configured to render an interface comprising the first post and the modified suggested keyword query, wherein the first post is displayed in association with the modified suggested keyword query.

20. The method of claim 1, wherein the first post has been tagged with one or more topics, and wherein the topic-score for each topic associated with the extracted keywords is further based on a relevance of the topic to the one or more topics tagged in the first post.

21. The media of claim 18, wherein the software is further operable when executed to:

accessing a social graph comprising a plurality of nodes and a plurality of edges connecting the nodes, each of the edges between two of the nodes representing a single degree of separation between them, the nodes comprising;

a first node corresponding to the first user; and

a plurality of second nodes corresponding to a plurality of posts associated with the online social network, respectively.

22. The media of claim 18, wherein the software, when executed to determine, for each extracted keyword, the one or more topics associated with the extracted keyword, is further operable to:

receive an electronic topic index comprising a set of objects, wherein each object represents a topic;

determine a context of the extracted keyword in the first post;

calculate a context-score for each object of the set of objects based on the determined context; and

determine whether there is an object having a context-score greater than a threshold context-score.

23. The media of claim 22, wherein the context of the extracted keyword in the first post is determined using a forward-backward algorithm to calculate a probability of the extracted keyword matching the topic based on adjacent n-grams that match the topic also being present.

24. The media of claim 18, wherein the relevance of the topic to the first post is based on an overall context of the first post.

25. The media of claim 18, wherein the topic-score is further based on a current popularity of the topic on the online social network.

26. The media of claim 18, wherein the topic-score is further based on a confidence level that a particular extracted keyword associated with the topic corresponds to the topic, the confidence level being based on a context in which the particular extracted keyword is used in the post.

27. The media of claim 18, wherein the software, when executed to extract the plurality of keywords, is further operable to filter out one or more n-grams from the content of the first post having a low relevance based on a TF-IDF analysis.

28. The media of claim 18, wherein the software is further operable when executed to remove a name of an author of the first post from the extracted keywords.

29. The media of claim 18, wherein the suggested keyword query comprises a fixed number of extracted keywords.

30. The media of claim 18, wherein the software is further operable when executed to send instructions for display indicating that the modified suggested keyword query corresponding to the first post can be used to find the first post again or to find related posts.

31. The media of claim 18, wherein the modified suggested keyword query is displayed adjacent to the first post.

32. The media of claim 18, wherein the extracted keywords in the modified suggested keyword query are displayed in an order based on a relative position of the extracted keywords in the content of the first post.

33. The media of claim 18, wherein the suggested keyword is a link that is selectable to execute a search query comprising the modified suggested keyword query.

34. The media of claim 18, wherein the software is further operable when executed to verify that the suggested keyword query retrieves the first post in response to executing the modified suggested keyword query.

35. The media of claim 18, wherein the software is further operable when executed to:

receive, from the client device of the first user, a search query comprising the modified suggested keyword query; and

generate one or more search results comprising references to one or more posts of the online social network matching the modified suggested keyword query, wherein at least one of the search results comprises a reference to the first post.

36. The media of claim 18, wherein the software is further operable when executed to:

receive, from the client device of the first user, a search query comprising the modified suggested keyword query; and

generate one or more search results comprising references to one or more second posts of the online social network matching the modified suggested keyword query, wherein at least one of the second posts is a related post.