Social message monitoring method and apparatus

• US 10,250,550 B2
• Filed: 05/05/2016
• Issued: 04/02/2019
• Est. Priority Date: 04/28/2014
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Chen, Lei, Zhou, Xiangmin, He, Xiuqiang
• Primary Examiner(s): Huq, Farzana B
• Application Number:
  US15/147,684
• Publication Number:
  US 20160248724A1
• Time in Patent Office: 1,062 Days
• Field of Search: 709206, 709224, 709226
• CPC Class Codes: G06F 16/00 Information retrieval; Database structures therefor; File system structures therefor
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks
• CPC Class Codes: G06Q 50/01 Social networking
• CPC Class Codes: H04L 51/234 for tracking messages
• CPC Class Codes: H04L 51/52 for supporting social networking services

Abstract

A social message monitoring method is implemented by receiving, from a social network server, a social message, and obtaining a theme probability vector of the social message, comparing the theme probability vector of the social message with a theme probability vector of each representative message to obtain a theme similarity, and acquiring a similarity between the social message and each representative message according to the theme similarity, and saving the social message in a message class that contains a representative message most similar to the social message, and outputting the message class to a social network client when a quantity of social messages in the message class reaches a first threshold or themes of social messages in the message class are consistent.

Claims

1. A social message monitoring method, comprising:

receiving a social message;

performing theme modeling analysis on the social message to obtain a theme probability vector of the social message;

comparing the theme probability vector of the social message with a theme probability vector of a representative message to obtain a theme similarity;

comparing a user of the social message with a user of each representative message to obtain a user similarity between the social message and each representative message;

acquiring a similarity between the social message and the representative message according to the theme similarity, the representative message being a representative social message in a message class;

saving the social message in a message class containing a representative message most similar to the social message, the representative message from each message class being selected by performing weighted averaging on theme probability vectors of all social messages in each message class to obtain a theme probability vector of a representative message of each message class, and when a new social message is added to the message class, selecting, from each message class, a social message matching the theme probability vector of the representative message of the message class as the representative message, and performing weighted averaging again to calculate the theme probability vector of the representative message of the message class; and

outputting the message class to a social network client when a quantity of social messages in the message class reaches a first threshold or themes of social messages in the message class are consistent.

2. The method of claim 1, further comprising:

creating a new message class when the similarity between the social message and each of the representative message is not greater than a preset first similarity threshold; and

saving the social message in the new message class when the similarity between the social message and each of the representative message is not greater than the preset first similarity threshold.

3. The method of claim 1, wherein receiving the social message, and performing the theme modeling analysis on the social message to obtain the theme probability vector of the social message comprises:

receiving the social message;

acquiring a content parameter of the social message, a temporal parameter of the social message, and a spatial parameter of the social message; and

performing the theme modeling analysis on the content parameter of the social message, the temporal parameter of the social message, and the spatial parameter of the social message using a latent Dirichlet allocation (LDA) algorithm, to obtain the theme probability vector of the social message.

4. The method of claim 2, wherein receiving the social message, and performing the theme modeling analysis on the social message to obtain the theme probability vector of the social message comprises:

receiving the social message;

acquiring a content parameter of the social message, a temporal parameter of the social message, and a spatial parameter of the social message; and

performing the theme modeling analysis on the content parameter of the social message, the temporal parameter of the social message, and the spatial parameter of the social message using a latent Dirichlet allocation (LDA) algorithm, to obtain the theme probability vector of the social message.

5. The method of claim 1, further comprising:

comparing the theme probability vector of the social message with a theme probability vector of each representative message to obtain a theme similarity, and

the similarity between the social message and each representative message being acquired according to the theme similarity by;

calculating a relative entropy between the theme probability vector of the social message and the theme probability vector of each representative message to obtain a theme message similarity between the social message and each representative message; and

performing weighted summation on the theme message similarity and the user similarity to obtain the similarity between the social message and each representative message.

6. The method of claim 1, further comprising performing a search, insert, or delete operation on a social message in any message class.

7. The method of claim 1, wherein before performing the theme modeling analysis on the social message to obtain the theme probability vector of the social message, the method further comprises:

acquiring a field to which the social message belongs; and

removing the social message to obtain a filtered social message when the field to which the social message belongs is not a preset field, and

the theme modeling analysis on the social message is performed to obtain a theme probability vector of the social message by performing the theme modeling analysis on the filtered social message to obtain the theme probability vector of the social message.

8. The method of claim 2, wherein before performing the theme modeling analysis on the social message to obtain the theme probability vector of the social message, the method further comprises:

acquiring a field to which the social message belongs; and

removing the social message to obtain a filtered social message when the field to which the social message belongs is not a preset field, and

the theme modeling analysis on the social message is performed to obtain a theme probability vector of the social message by performing the theme modeling analysis on the filtered social message to obtain the theme probability vector of the social message.

9. A social network server configured to monitor a social message, comprising:

a memory comprising instructions; and

a processor coupled to the memory and configured to execute the instructions to;

receive a social message using a communications interface;

perform theme modeling analysis on the social message to obtain a theme probability vector of the social message;

compare the theme probability vector of the social message with the theme probability vector of a representative message to obtain a theme similarity;

compare a user of the social message with a user of each representative message to obtain a user similarity between the social message and each representative message;

acquire a similarity between the social message and the representative message according to the theme similarity, the representative message being a representative social message in a message class;

save the social message in a message class that contains a representative message most similar to the social message, the similarity being obtained by measurement, the representative message from each message class being selected by performing weighted averaging on theme probability vectors of all social messages in each message class to obtain a theme probability vector of a representative message of each message class, and when a new social message is added to the message class, selecting, from each message class, a social message matching the theme probability vector of the representative message of the message class as the representative message, and performing weighted averaging again to calculate the theme probability vector of the representative message of the message class; and

output the message class to a social network client when a quantity of social messages in the message class reaches a first threshold or themes of social messages in the message class are consistent.

10. The social network server of claim 9, wherein the processor is further configured to:

create a new message class when the similarity between the social message and each of the representative messages is not greater than a preset first similarity threshold; and

save the social message in the new message class when the similarity between the social message and each of the representative messages is not greater than the preset first similarity threshold.

11. The social network server of claim 9, wherein the processor is further configured to:

receive the social message;

acquire a content parameter of the social message, a temporal parameter of the social message, and a spatial parameter of the social message; and

perform the theme modeling analysis on the content parameter of the social message, the temporal parameter of the social message, and the spatial parameter of the social message using a latent Dirichlet allocation (LDA) algorithm, to obtain the theme probability vector of the social message.

12. The social network server of claim 10, wherein the processor is further configured to:

receive the social message;

acquire a content parameter of the social message, a temporal parameter of the social message, and a spatial parameter of the social message; and

perform the theme modeling analysis on the content parameter of the social message, the temporal parameter of the social message, and the spatial parameter of the social message using a latent Dirichiet allocation (LDA) algorithm, to obtain the theme probability vector of the social message.

13. The social network server of claim 9, wherein the processor is further configured to:

calculate a relative entropy between the theme probability vector of the social message and the theme probability vector of each representative message to obtain a theme message similarity between the social message and each representative message; and

perform weighted summation on the theme message similarity and the user similarity, to obtain the similarity between the social message and each representative message.

14. The social network server of claim 9, wherein the processor is further configured to perform a search, insert, or delete operation on a social message in any message class.

15. The social network server of claim 9, wherein the processor is further configured to:

acquire a field to which the social message belongs;

remove the social Message to obtain a filtered social message when the field to which the social message belongs is not a preset field; and

perform the theme modeling analysis on the filtered social message to obtain the theme probability vector of the social message.

16. The social network server of claim 10, wherein the processor is further configured to;

acquire a field to which the social message belongs;

remove the social message to obtain a filtered social message when the field to which the social message belongs is not a preset field; and

perform the theme modeling analysis on the filtered social message to obtain the theme probability vector of the social message.