Connecting people based on content and relational distance

Patent metadata

Field	Value
Publication number	US-10002127-B2
Application number	US-201414308124-A
Country	US
Kind code	B2
Filing date	Jun 18, 2014
Priority date	Jan 17, 2014
Publication date	Jun 19, 2018
Grant date	Jun 19, 2018

Abstract

Technologies for establishing connections between or among people based at least in part on semantic distance and relational distance include one or more computing devices that analyze content created by computing device users, perform content clustering on the content, determine relational distances between connection candidates, and generate connection recommendations based at least in part on the content clustering and the relational distances.

First claim

The invention claimed is:

1. A computing device for generating a connection recommendation, the computing device comprising, for a computing device user in a set of computing device users:
a corpus development module to include user-generated electronic content in a corpus;
a lexical space builder module to select, from the corpus, a subset of low polysemy lexical items having a low corpus frequency and a low language frequency;
a clustering module to perform content clustering on the selected subset and subsets of extracted lexical items of other users in the set of users; and
a connection generator module to (i) identify a set of connection candidates comprising one or more of the other users in the set of users based at least in part on the content clustering; (ii) assign a weight to each connection candidate based on a relational distance between the corresponding connection candidate and the user; (iii) compare the weight of each connection candidate to a reference threshold; (iv) remove, from the identified set of connection candidates, those connection candidates that have an assigned weight less than the reference threshold; (v) present a connection recommendation identifying at least one of the connection candidates remaining in the set of connection candidates after those connection candidates having assigned weights less than the reference threshold are removed from the set of connection candidates, (vi) receive context data obtained by a sensor of the computing device, and (vii) modify the connection recommendation based at least in part on the context data, wherein the context data includes a current geographic location of the user.

2. The computing device of claim 1 , wherein the corpus development module is to select the user-generated electronic content for inclusion in the corpus based at least in part on the context data.

3. The computing device of claim 1 , wherein the connection generator module is to modify the connection recommendation based at least in part on a characteristic of the content, and the characteristic of the content is algorithmically determined by one or more of a topic analysis and a stylometric analysis.

4. The computing device of claim 1 , wherein the corpus development module is to select the user-generated electronic content based at least in part on a characteristic of the content, wherein the characteristic of the content is algorithmically determined by one or more of a topic analysis and a stylometric analysis.

5. The computing device of claim 1 , wherein the lexical space builder module is to build a multidimensional space comprises a number of dimensions corresponding to a number of words in all of the corpora of the computing device users, and each of the dimensions of the multidimensional space represents a lexical item existing in at least one of the corpora.

6. The computing device of claim 1 , wherein the lexical space builder module is to assign weights to the extracted lexical items according to the language frequency of the extracted lexical items.

7. The computing device of claim 1 , comprising a social network identifier module to, for each of the computing device users, identify a social network of the computing device user, wherein the connection generator module obtains data indicating the relational distances between the computing device user and the connection candidate from the identified social network.

8. The computing device of claim 7 , comprising a user profile generator module to develop a user profile comprising data relating to the subset of extracted lexical items and the identified social network, wherein the connection generator module is to generate the connection recommendation based at least in part on the user profile.

9. The computing device of claim 1 , wherein the user-generated electronic content comprises natural language input.

10. The computing device of claim 1 , wherein the clustering module is to compute numerical distances between the subsets of extracted lexical items of pairs of computing device users.

11. The computing device of claim 10 , wherein the connection generator module is to generate the connection recommendation if the computed numerical distance between the subsets of extracted lexical items of a pair of computing device users satisfies the reference threshold.

12. One or more non-transitory, machine readable storage media comprising a plurality of instructions stored thereon that in response to being executed result in a computing device:
including user-generated electronic content in a corpus;
selecting, from the corpus, a subset of low polysemy lexical items having a low corpus frequency and a low language frequency;
performing content clustering on the selected subset and subsets of extracted lexical items of other users in the set of users;
identifying a set of connection candidates comprising one or more of the other users in the set of users based at least in part on the content clustering;
assigning a weight to each connection candidate based on a relational distance between the corresponding connection candidate and the user;
comparing the weight of each connection candidate to a reference threshold;
removing, from the identified set of connection candidates, those connection candidates that have an assigned weight less than the reference threshold;
presenting a connection recommendation identifying at least one of the connection candidates remaining in the set of connection candidates after those connection candidates having assigned weights less than the reference threshold are removed from the set of connection candidates;
receiving context data obtained by a sensor of the computing device; and
modifying the connection recommendation based at least in part on the context data, wherein the context data includes a current geographic location of the user.

13. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device selecting the user-generated electronic content based at least in part on the context data.

14. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device modifying the connection recommendation based at least in part on a characteristic of the content, wherein the characteristic of the content is algorithmically determined by one or more of a topic analysis and a stylometric analysis.

15. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device selecting the user-generated electronic content based at least in part on a characteristic of the content, wherein the characteristic of the content is algorithmically determined by one or more of a topic analysis and a stylometric analysis.

16. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device assigning weights to the extracted lexical items according to the language frequency of the extracted lexical items.

17. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device identifying a social network of the computing device user and obtaining data indicating the relational distances between the computing device user and the connection candidate from the identified social network.

18. The one or more non-transitory, machine readable storage media of claim 12 , wherein the instructions result in the computing device de

Assignees

• Beckwith Richard T
• Anderson Kenneth T
• Bezaitis Maria
• Wouhaybi Rita H
• Intel Corp

Inventors

• Beckwith Richard T
• Anderson Kenneth T
• Bezaitis Maria
• Wouhaybi Rita H

Classifications

• G06F40/284 Primary

  Lexical analysis, e.g. tokenisation or collocates · CPC title

• G06F16/3344

  using natural language analysis · CPC title

• G06F16/248

  Presentation of query results · CPC title

• G06F16/313

  Selection or weighting of terms for indexing · CPC title

• G06F16/951

  Indexing; Web crawling techniques · CPC title