Coefficients Attribution for Different Objects Based on Natural Language Processing

• US 20130041653A1
• Filed: 08/12/2011
• Published: 02/14/2013
• Est. Priority Date: 08/12/2011
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Tseng, Erick
• Granted Patent: US 9,530,167 B2
• Time in Patent Office: Days
• US Class Current: 704/9
• CPC Class Codes: G06F 16/90324 using system suggestions
• CPC Class Codes: G06F 16/904 Browsing; Visualisation therefor for navigating the web G06F16/954; browsing optimisation for the web G06F16/957
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/958 Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

In one embodiment, a system includes one or more computing systems that implement a social networking environment and is operable to parse users'"'"' actions that include free form text to determine and store objects and affinities contained in the text string through natural-language processing. The method comprises accessing a text string, identifying objects and affinity declarations via natural-language processing, assessing the combination of objects and context data to determine an instance of a broader concept, and determining an affinity coefficient through a natural-language processing dictionary. Once a database of stored instances and affinities has been generated and stored, it may be leveraged to push suggestions to members of the social network to enhance their social networking experience.

Claims

1. A method comprising, by one more computing systems:

receiving an input from a user, wherein the input comprises free-form text;

determining, through application of natural-language processing of the free-form text, one or more objects associated with the input;

determining, through application of natural-language processing of the free-form text, one or more affinity declarations associated with the one or more objects;

determining, from the one or more objects, an instance of a concept;

determining an affinity coefficient between the user and the instance based on the affinity declaration; and

storing the affinity coefficient in a data store in association with the user and the instance.

2. The method of claim 1, wherein the input also comprises context data.

3. The method of claim 2, wherein the context data comprises a location.

4. The method of claim 2, wherein the context data comprises a time.

5. The method of claim 2, wherein the context data comprises a check-in to a particular location.

6. The method of claim 2, wherein the context data is inferred based on an element that the user input acts on.

7. The method of claim 1, wherein the affinity coefficient may be a positive or negative numerical value.

8. The method of claim 7, wherein the natural-language processing utilizes a dictionary comprising crowd-sourced adjectives or objects.

9. The method of claim 8, wherein the affinity coefficient is based on the relative strength of the adjectives found in the affinity declaration.

10. The method of claim 1, further comprising:

determining whether the instance exists as a node in a social graph; and

if the instance does not exist in the social graph, creating a node in the social graph representing the determined instance.

11. The method of claim 10, wherein the node is hidden from users of a social network accessing the social graph.

12. The method of claim 11, wherein storing the affinity comprises creating an edge connecting a node representing the user and the node representing the instance:

13. The method of claim 1, wherein the user is represented by a first node in a social graph, and the instance is represented by a second node in the social graph, and wherein storing the affinity coefficient comprises creating an edge connecting the first and second nodes.

14. A method comprising, by one or more computing systems:

receiving an input from a user, the input comprising free-form text;

determining, through application of natural-language processing to the free-form text, one or more objects associated with the input;

searching a database of instances for instances substantially related to the determined one or more objects having a predetermined number of positive affinity coefficients associated with the instance; and

delivering a notification to the user recommending the instance.

15. The method of claim 14, wherein the positive affinity coefficients are above a predetermined numerical value.

16. A non-transitory, computer-readable media comprising instructions, the instructions operable, when executed, to:

receive an input from a user, wherein the input comprises free-form text;

determine, through application of natural-language processing of the free-form text, one or more objects associated with the input;

determine, through application of natural-language processing of the free-form text, one or more affinity declarations associated with the one or more objects;

determine, from the one or more objects, an instance of a concept;

determine an affinity coefficient between the user and the instance based on the affinity declaration; and

store the affinity coefficient in a data store in association with the user and the instance.

17. The media of claim 16, wherein the input also comprises context data.

18. The media of claim 17, wherein the context data comprises a location or time.

19. The media of claim 17, wherein the context data comprises a check-in to a particular location.

20. The media of claim 17, wherein the context data is inferred based on an element that the user input acts on.