Time and sentiment based messaging

• US 10,462,095 B2
• Filed: 01/10/2017
• Issued: 10/29/2019
• Est. Priority Date: 01/10/2017
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Abou Mahmoud, Alaa, Bastide, Paul R., Lu, Fang
• Primary Examiner(s): Keller, Michael A
• Assistant Examiner(s): Duong, Thao D
• Application Number:
  US15/402,365
• Publication Number:
  US 20180198749A1
• Time in Patent Office: 1,022 Days
• CPC Class Codes: H04L 51/52 for supporting social networking services
• CPC Class Codes: H04L 67/62 Establishing a time schedule for servicing the requests
• CPC Class Codes: H04W 4/21 for social networking applications

Abstract

In time and sentiment based messaging, a server obtains information for a set of messages from online social networks related to a specified object. The information includes at least a user identifier associated with each message, a time of each message, and content of each message. For each unique identifier, the server establishes an initial message from the set of messages related to the specified object, analyzes each message in the set of messages to determine a sentiment of each message, and builds a sentiment line time for each unique user identifier. The server then builds a time-based sentiment model related to the specified object by overlapping the sentiment time lines for each unique user identifier according to the initial message for each unique user identifier. A targeted message to a set of one or more users may then be generated using the time-based sentiment model.

Claims

1. A method for time and sentiment based messaging, comprising:

obtaining, by a server, information for a set of messages from online social networks related to a specified object, the information comprising at least a user identifier associated with each message, a time of each message, and content of each message;

for each unique identifier, establishing, by the server, an initial message from the set of messages related to the specified object;

analyzing, by the server, each message in the set of messages to determine a sentiment of each message toward the specified object;

building, by the server, a sentiment time line for each unique user identifier using the sentiment of each message toward the specified object;

building, by the server, a time-based sentiment model related to the specified object by overlapping the sentiment time lines for each unique user identifier according to the initial message for each unique user identifier;

identifying, by the server, a sentiment inflection point in the time-based sentiment model, the sentiment inflection point representing a change in the sentiment toward the specified object;

building, by the server, a new sentiment time line for an additional unique user identifier;

overlapping, by the server, the new sentiment time line with the time-based sentiment model according to an initial message for the additional unique user identifier;

predicting, by the server, a change in the sentiment related to the specified object by the additional unique user identifier based on the new sentiment time line, the time-based sentiment model, and the sentiment inflection point; and

generating a message targeting the predicted change in the sentiment related to the specified object by the additional unique user identifier.

2. The method of claim 1, wherein the establishing of the initial message from the set of messages related to the specified object for each unique identifier comprises:

identifying the unique user identifiers in the set of messages and organizing the set of messages according to the unique user identifiers;

normalizing the set of messages based on the time of each message in the set; and

for each unique user identifier, establishing the initial message.

3. The method of claim 1, wherein the time-based sentiment model relates to the specified object over a lifecycle of the specified object.

4. The method of claim 1, wherein the sentiment inflection point is a point in a lifecycle of the specified object when sentiments toward the specified object change.

5. A computer program product for time and sentiment based messaging, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to:

obtain information for a set of messages from online social networks related to a specified object, the information comprising at least a user identifier associated with each message, a time of each message, and content of each message;

for each unique identifier, establish an initial message from the set of messages related to the specified object;

analyze each message in the set of messages to determine a sentiment of each message toward the specified object;

build a sentiment time line for each unique user identifier using the sentiment of each message toward the specified object;

build a time-based sentiment model related to the specified object by overlapping the sentiment time lines for each unique user identifier according to the initial message for each unique user identifier;

identify a sentiment inflection point in the time-based sentiment model, the sentiment inflection point representing a change in the sentiment toward the specified object;

build a new sentiment time line for an additional unique user identifier;

overlap the new sentiment time line with the time-based sentiment model according to an initial message for the additional unique user identifier;

predict a change in the sentiment related to the specified object by the additional unique user identifier based on the new sentiment time line, the time-based sentiment model, and the sentiment inflection point; and

generate a message targeting the predicted change in the sentiment related to the specified object by the additional unique user identifier.

6. The computer program product of claim 5, wherein the establishing of the initial message from the set of messages related to the specified object for each unique identifier comprises:

identify the unique user identifiers in the set of messages and organizing the set of messages according to the unique user identifiers;

normalize the set of messages based on the time of each message in the set; and

for each unique user identifier, establish the initial message.

7. The computer program product of claim 5, wherein the time-based sentiment model relates to the specified object over a lifecycle of the specified object.

8. The computer program product of claim 5, wherein the sentiment inflection point is a point in a lifecycle of the specified object when sentiments toward the specified object change.

9. A system, comprising:

a processor; and

a computer readable storage medium having program instructions embodied therewith, the program instructions executable by the processor to cause the processor to;

obtain information for a set of messages from online social networks related to a specified object, the information comprising at least a user identifier associated with each message, a time of each message, and content of each message;

for each unique identifier, establish an initial message from the set of messages related to the specified object;

analyze each message in the set of messages to determine a sentiment of each message toward the specified object;

build a sentiment time line for each unique user identifier using the sentiment of each message toward the specified object;

build a time-based sentiment model related to the specified object by overlapping the sentiment time lines for each unique user identifier according to the initial message for each unique user identifier;

identify a sentiment inflection point in the time-based sentiment model, the sentiment inflection point representing a change in the sentiment toward the specified object;

build a new sentiment time line for an additional unique user identifier;

overlap the new sentiment time line with the time-based sentiment model according to an initial message for the additional unique user identifier;

predict a change in the sentiment related to the specified object by the additional unique user identifier based on the new sentiment time line, the time-based sentiment model, and the sentiment inflection point; and

generate a message targeting the predicted change in the sentiment related to the specified object by the additional unique user identifier.

10. The system of claim 9, wherein the time-based sentiment model relates to the specified object over a lifecycle of the specified object.

11. The system of claim 9, wherein the sentiment inflection point is a point in a lifecycle of the specified object when sentiments toward the specified object change.