Message Generation for Ranked Users Based on User Interaction Probabilities

Patent metadata

Field	Value
Publication number	US-2020013092-A1
Application number	US-201816028533-A
Country	US
Kind code	A1
Filing date	Jul 6, 2018
Priority date	Jul 6, 2018
Publication date	Jan 9, 2020
Grant date	—

Abstract

An approach is provided in which a system determines a set of message properties corresponding to a set of products based on product data analysis. The system then identifies a set of user properties of users based on analyzing social media data corresponding to the product data. Next, the system identifies a set of candidate customers from the set of users based on analyzing the set of user properties against the set of product data. In turn, the system generates a set of target messages that are tailored to a combination of candidate customer properties corresponding to the set of candidate customers and the product data.

First claim

1 . A method implemented by an information handling system that includes a memory and a processor, the method comprising:
determining a set of brand personalities corresponding to a set of products based on analyzing brand data corresponding to the set of products; identifying a set of user personalities of a set of users based on analyzing social media data corresponding to the set of products generated by the set of users; identifying a set of candidate customers from the set of users based on analyzing the set of user personalities against the set of brand personalities, wherein the set of candidate customers correspond to a set of candidate customer personalities comprised in the set of user personalities; and creating a set of target messages wherein each target message is tailored to a combination of one of the candidate customer personalities and one of the brand personalities.

2 . The method of claim 1 wherein the social media data comprises a plurality of user interactions written by the set of users and corresponding to a set of current messages, the method further comprising:
assigning a message type to each of the set of current messages, resulting in a set of message types;
creating, for each of the plurality of user interactions, an interaction vector that comprises one of the plurality of user interactions, a corresponding one of the set of message types, a corresponding one of the set of user personalities, and a corresponding one of the set of brand personalities, resulting in a plurality of interaction vectors; and
generating, based on the plurality of interaction vectors, a user interaction preferences matrix, a user personality matrix and a brand personality matrix, wherein the user interaction preferences matrix represents the plurality of user interactions, the user personality matrix represents the set of user personalities, and the brand personality matrix represents the set of brand personalities.

3 . The method of claim 2 further comprising:
computing a bridging matrix that indicates a correlation between the user personality matrix and the brand personality matrix relative to the user interaction preferences matrix; and
selecting the set of candidate customers from the set of users based on the bridging matrix.

4 . The method of claim 2 wherein the message type is selected from the group consisting of a relationship maintenance message type, a new product release message type, a customer engagement enhancement message type, and a promotion/sales message type.

5 . The method of claim 1 further comprising:
training the information handling system using a set of training data comprising different social media data corresponding to a different set of products that are similar to the set of products;
generating a set of sample messages using the trained information handling system; and
generating a set of brand personality scores for each of the sample messages based on comparing the set of sample messages against the set of brand personalities.

6 . The method of claim 5 further comprising:
generating a set of customer personalization scores for each of the set of sample messages based on comparing the set of sample messages against a set of verbal expressions comprised in the social media data and written by the set of candidate customers; and
aggregating each one of the set of customer personalization scores with each corresponding one of the set of brand personality scores, resulting in a set of aggregation scores each corresponding to one of the set of sample messages that optimize a tradeoff between the set customer personalization scores and the set of brand personality scores.

7 . The method of claim 6 further comprising:
selecting, based on the set of aggregation scores, the et of target messages from the set of sample messages; and
sending the set of target messages to the set of corresponding candidate customers.

8 . An information handling system comprising:
one or more processors; a memory coupled to at least one of the processors; a set of computer program instructions stored in the memory and executed by at least one of the processors in order to perform actions of:
determining a set of brand personalities corresponding to a set of products based on analyzing brand data corresponding to the set of products;
identifying a set of user personalities of a set of users based on analyzing social media data corresponding to the set of products generated by the set of users;
identifying a set of candidate customers from the set of users based on analyzing the set of user personalities against the set of brand personalities, wherein the set of candidate customers correspond to a set of candidate customer personalities comprised in the set of user personalities; and
creating a set of target messages wherein each target message is tailored to a combination of one of the candidate customer personalities and one of the brand personalities.

9 . The information handling system of claim 8 wherein the social media data comprises a plurality of user interactions written by the set of users and corresponding to a set of current messages, and wherein the processors perform additional actions comprising:
assigning an message type to each of the set of current messages, resulting in a set of message types;
creating, for each of the plurality of user interactions, an interaction vector that comprises one of the plurality of user interactions, a corresponding one of the set of message types, a corresponding one of the set of user personalities, and a corresponding one of the set of brand personalities, resulting in a plurality of interaction vectors; and
generating, based on the plurality of interaction vectors, a user interaction preferences matrix, a user personality matrix and a brand personality matrix, wherein the user interaction preferences matrix represents the plurality of user interactions, the user personality matrix represents the set of user personalities, and the brand personality matrix represents the set of brand personalities.

10 . The information handling system of claim 9 wherein the processors perform additional actions comprising:
computing a bridging matrix that indicates a correlation between the user personality matrix and the brand personality matrix relative to the user interaction preferences matrix; and
selecting the set of candidate customers from the set of users based on the bridging matrix.

11 . The information handling system of claim 9 wherein the message type is selected from the group consisting of a relationship maintenance message type, a new product release message type, a customer engagement enhancement message type, and a promotion/sales message type.

12 . The information handling system of claim 8 wherein the processors perform additional actions comprising:
training the information handling system using a set of training data comprising different social media data corresponding to a different set of products that are similar to the set of products;
generating a set of sample messages using the trained information handling system; and
generating a set of brand personality scores for each of the sample messages based on comparing the set of sample messages against the set of brand personalities.

13 . The information handling system of claim 12 wherein the processors perform additional actions comprising:
generating a set of customer personalization scores for each of the set of sample messages based on comparing the set of sample messages against a set of verbal expressions comprised in the social media data and written by the set of candidate

Assignees

• IBM

Inventors

• Liu Zhe
• Zhang Mengdi
• Xu Anbang
• Webster April L
• Mahmud Jalal U
• Zimmerman Thomas G
• Akkiraju Rama Kalyani T
• Sinha Vibha S

Classifications

• G06N20/00

  Machine learning · CPC title

• G06Q10/40

  Business processes related to social networking or social networking services · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06Q30/0271 Primary

  Personalized advertisement · CPC title

• G06Q50/01

  Physics · mapped topic