Method and apparatus for providing a response to an input post on a social page of a brand

• US 10,269,080 B2
• Filed: 11/25/2014
• Issued: 04/23/2019
• Est. Priority Date: 11/25/2014
• Status: Active Grant

• Current Assignee: Adobe Inc.
• Original Assignee: Adobe Inc.
• Inventors: Dhawan, Anmol, Chang, Walter Wei-Tuh, Duggal, Ashish, Soni, Sachin
• Primary Examiner(s): Dailey, Thomas J
• Assistant Examiner(s): Aung, Aye M
• Application Number:
  US14/553,292
• Publication Number:
  US 20160148325A1
• Time in Patent Office: 1,610 Days
• Field of Search: 709217
• CPC Class Codes: G06Q 30/0269 based on user profile or attribute
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

A method for providing a response to an input post on a social page of a brand is provided. The input post is detected upon posting of the input post on the social page of the brand. The social page is present on a social channel. An inquiry regarding the brand is identified from content of the input post. At least one social post is determined from already posted posts on one or more social channels based on the inquiry. The at least one social post is associated with the brand. A response post is created using the at least one social post. The response post addresses the inquiry. The response post is then posted on the social page of the social channel as a reply to the input post. An apparatus for performing the method as described herein is also provided.

Claims

1. A method comprising:

detecting an input post, by an apparatus comprising at least one processor, upon posting of the input post on a social page of a brand, the social page being present on a social channel;

identifying, by the apparatus, an inquiry regarding the brand from the input post;

extracting, by the apparatus, one or more keywords from the inquiry;

determining, by the apparatus, semantically equivalent terms for the extracted one or more keywords;

generating, by the apparatus, a feature vector for the input post by using the one or more keywords and the semantically equivalent terms;

determining, by the apparatus, at least one social post from already posted posts on one or more social channels, the at least one social post associated with the brand;

accessing a database storing a mapping between the feature vector for the input post and one or more posts previously determined to be relevant to the feature vector for the input post;

comparing a post time associated with the at least one social post and a latest mapping time associated with the mapping stored in the database; and

based on determining that the post time associated with the at least one social post is earlier than the latest mapping time associated with the mapping;

creating, by the apparatus, based on a similarity measure between a feature vector for the at least one social post and the feature vector for the input post, a response post using the at least one social post; and

causing to post the response post, by the apparatus, on the social page of the social channel as a reply to the input post.

2. The method as claimed in claim 1, wherein extracting, by the apparatus, the one or more keywords from the inquiry comprises:

using a part of speech tagger trained on content associated with the brand.

3. The method as claimed in claim 1, wherein determining the at least one social post comprises:

identifying a social post having an associated feature vector within a predefined range of the feature vector for the input post as the at least one social post.

4. The method as claimed in claim 3, wherein determining the at least one social post comprises:

crawling one or more social posts on the one or more social channels; and

performing sentiment analysis on the one or more social posts to filter out social posts that have negative sentiments.

5. The method as claimed in claim 3, wherein determining the at least one social post comprises:

crawling one or more social posts on the one or more social channels;

generating feature vectors for the one or more social posts; and

determining a similarity score between each feature vector of the feature vectors for the one or more social posts and the feature vector for the input post.

6. The method as claimed in claim 5 and further comprising:

storing the similarity score between each feature vector of the feature vectors for the one or more social posts and the feature vector for the input post for further use.

7. The method as claimed in claim 1, wherein creating the response post comprises:

recommending the at least one social post to a marketer of the brand;

receiving an input from the marketer; and

generating the response post based on the input of the marketer and based on content of the at least one social post.

8. The method as claimed in claim 1, wherein creating the response post comprises:

extracting information from links present in the at least one social post; and

generating the response post based on the information.

9. The method as claimed in claim 1, wherein the at least one social post comprises at least one of:

a social post posted by a marketer of the brand on the one or more social channels, the one or more social channels comprising the social channel via which the input post is received;

a social post posted by a user on the one or more channels;

or

a link posted by the user on the one or more channels.

10. The method as claimed in claim 1, further comprising based on determining that the post time associated with the at least one social post is later than the latest mapping time associated with the mapping:

determining, by the apparatus, a feature vector for a second social post based on one or more keywords included in the second social post;

determining, by the apparatus, the similarity measure between the feature vector for second social post and the feature vector for the input post;

based on the similarity measure, determining, by the apparatus, whether the at least one social post is spam using a machine learning algorithm;

based on determining that the at least one social post is not spam, mapping the second social post to the feature vector for the input post in the database;

creating, by the apparatus, based on the similarity measure, the response post using the second social post; and

causing to post the response post, by the apparatus, on the social page of the social channel as a reply to the input post.

11. A non-transitory computer readable storage medium comprising instructions that, when executed by at least one processor, cause a computing device to:

detect an input post upon posting of the input post on a social page of a brand, the social page being present on a social channel;

identify an inquiry regarding the brand from content of the input post;

extract one or more keywords from the inquiry;

determine semantically equivalent terms for the extracted one or more keywords;

generate a feature vector for the input post by using the one or more keywords and the semantically equivalent terms;

determine at least one social post from already posted posts on one or more social channels, the at least one social post associated with the brand;

access a database storing a mapping between the feature vector for the input post and one or more posts previously determined to be relevant to the feature vector for the input post;

compare a time associated with the at least one social post and a latest mapping time associated with the mapping stored in the database; and

based on determining that the post time associated with the at least one social post is earlier than the latest mapping time associated with the mapping;

create, based on a similarity measure between a feature vector for the at least one social post and the feature vector for the input post, a response post using the at least one social post; and

cause to post the response post on the social page of the social channel as a reply to the input post.

12. The non-transitory computer readable storage medium as claimed in claim 11, wherein the instructions, when executed by the at least one processor, cause the computing device to determine the at least one social post by:

identifying a social post having an associated feature vector within a predefined range of the feature vector for the input post as the at least one social post.

13. The non-transitory computer readable storage medium as claimed in claim 12, wherein the instructions, when executed by the at least one processor, cause the computing device to determine the at least one social post by:

crawling one or more social posts on the one or more social channels;

generating feature vectors for the one or more social posts; and

determining a similarity score between each feature vector of the feature vectors for the one or more social posts and the feature vector for the input post.

14. The non-transitory computer readable storage medium as claimed in claim 11, wherein the instructions, when executed by the at least one processor, cause the computing device to create the response post by:

using content of the at least one social post as it is.

15. The non-transitory computer readable storage medium as claimed in claim 11, wherein the instructions, when executed by the at least one processor, cause the computing device to create the response post by:

recommending the at least one social post to a marketer of the brand;

receiving an input from the marketer; and

generating the response post based on the input of the marketer and based on content of the at least one social post.

16. The non-transitory computer readable storage medium as claimed in claim 11, wherein the instructions, when executed by the at least one processor, cause the computing device to create the response post by:

extracting information from links present in the at least one social post; and

generating the response post based on the information.

17. The non-transitory computer readable storage medium as claimed in claim 11, wherein the instructions, when executed by the at least one processor, cause the computing device to cause to post the response post by:

posting the response post on the social page in proximity to the input post.

18. An apparatus comprising:

a memory comprising computer-readable instructions; and

one or more processors coupled to the memory, wherein the computer-readable instructions, when executed by the one or more processors, cause the apparatus to;

detect an input post upon posting of the input post on a social page of a brand, the social page being present on a social channel;

identify an inquiry regarding the brand from content of the input post;

extract one or more keywords from the inquiry;

determine semantically equivalent terms for the extracted one or more keywords;

generate a feature vector for the input post by using the one or more keywords and the semantically equivalent terms;

determine at least one social post from already posted posts on one or more social channels, the at least one social post associated with the brand;

access a database storing a mapping between the feature vector for the input post and one or more posts previously determined to be relevant to the feature vector for the input post;

compare a post time associated with the at least one social post and a latest mapping time associated with the mapping stored in the database; and

based on determining that the post time associated with the at least one social post is earlier than the latest mapping time associated with the mapping;

create, based on a similarity measure between a feature vector for the at least one social post and the feature vector for the input post, a response post using the at least one social post; and

cause to post the response post, on the social page of the social channel, as a reply to the input post.

19. The apparatus as claimed in claim 18, wherein the instructions, when executed by the one or more processors, cause the apparatus to determine the at least one social post by:

identifying a social post having an associated feature vector within a predefined range of the feature vector for the input post as the at least one social post.

20. The apparatus as claimed in claim 18, wherein the at least one social post comprises at least one of:

a social post posted by a marketer of the brand on the one or more social channels, the one or more social channels comprising the social channel via which the input post is received;

a social post posted by a user on the one or more channels;

or

a link posted by the user on the one or more channels.