INFORMATION RECOMMENDATION METHOD AND APPARATUS IN SOCIAL MEDIA

• US 20160117595A1
• Filed: 01/08/2016
• Published: 04/28/2016
• Est. Priority Date: 07/11/2013
• 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: Zhang, Jie, Zhang, Yibo, Li, Hang
• Granted Patent: US 10,810,499 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 50/01 Social networking

Abstract

The present invention provides an information recommendation method and apparatus in a social media. An off-line procedure includes: determining a point of interest of a target user; selecting information related to the point of interest as an annotated corpus; and training an interest classification model of the target user by using the annotated corpus as a training sample. An on-line procedure includes: inputting to-be-recommended information to the interest classification model of the target user, so as to determine whether the to-be-recommended information tallies with an interest of the target user; and if the to-be-recommended information tallies with the interest of the target user, recommending the to-be-recommended information to the target user. According to the present invention, an effect of information recommendation in a social media can be improved.

Claims

1. An information recommendation method in a social media, the method comprising:

in an off-line procedure,

determining a point of interest of a target user from historical operation information that is of the target user and in the social media,

selecting information related to the point of interest from historical operation information that is of a quality user related to the point of interest and in the social media, as an annotated corpus, and

training an interest classification model of the target user by using the annotated corpus as a training sample; and

in an on-line procedure,

acquiring to-be-recommended information of the social media,

inputting the to-be-recommended information to the interest classification model of the target user, so as to determine whether the to-be-recommended information tallies with an interest of the target user, and

if the to-be-recommended information tallies with the interest of the target user, recommending the to-be-recommended information to the target user.

2. The method according to claim 1, wherein the historical operation information comprises at least one type of the following information:

posted information, favorited information, reposted information, comment information, and an annotated tag.

3. The method according to claim 2, wherein determining a point of interest of a target user from historical operation information that is of the target user and in the social media, comprises:

determining the point of interest of the user by using a tag that is annotated in the social media by the target user.

4. The method according to claim 2, wherein determining a point of interest of a target user from historical operation information that is of the target user and in the social media, comprises:

after word segmentation and stop word removal are separately performed on at least one type of the following information;

posted information, favorited information, reposted information, and comment information of the target user, performing semantic clustering on an obtained word, and determining the point of interest of the target user by using a clustering result.

5. The method according to claim 1, wherein a method for determining the quality user related to the point of interest comprises:

selecting a user whose score meets a preset requirement from a user who has an association relationship with a predetermined seed user related to the point of interest, as the quality user, wherein the score is determined according to one of or any combination of the following factors;

a degree of content matching between information that is posted, reposted, or self-described by the user and the point of interest, an activity degree of the user, impact of the user, and a spread range of user information.

6. The method according to claim 1, wherein selecting information related to the point of interest from historical operation information that is of a quality user related to the point of interest and in the social media, as an annotated corpus, comprises:

computing semantic similarity between the point of interest and each piece of information in the historical operation information of the quality user corresponding to the point of interest, and selecting information whose semantic similarity meets a preset similarity requirement as the information related to the point of interest.

7. The method according to claim 1, wherein when the interest classification model is trained, a used feature comprises a keyword of information.

8. The method according to claim 7, wherein when the interest classification model is trained, the used feature further comprises one of or any combination of the following factors:

an information poster, a person followed by the information poster or a follower of the information poster, a quantity of comments on information, a quantity of reposts of information, and a quantity of favorites of information.

9. The method according to claim 1, wherein the to-be-recommended information is real-time information, notification information, or promotion information.

10. A computer system, comprising:

a non-transitory memory and a processor, wherein the non-transitory memory stores instructions which when executed by the processor cause the processor to;

in an off-line procedure,

determine a point of interest of a target user from historical operation information that is of the target user and in the social media,

select information related to the point of interest from historical operation information that is of a quality user related to the point of interest and in the social media, as an annotated corpus, and

train an interest classification model of the target user by using the annotated corpus as a training sample; and

in an on-line procedure,

acquire to-be-recommended information of the social media,

input the to-be-recommended information to the interest classification model of the target user, so as to determine whether the to-be-recommended information tallies with an interest of the target user, and

if the to-be-recommended information tallies with the interest of the target user, recommend the to-be-recommended information to the target user.

11. The computer system according to claim 10, wherein the historical operation information comprises at least one type of the following information:

posted information, favorited information, reposted information, comment information, and an annotated tag.

12. The computer system according to claim 11, wherein the instructions further cause the processor to:

determine the point of interest of the user by using a tag that is annotated in the social media by the target user.

13. The computer system according to claim 11, wherein the instructions further cause the processor to:

perform semantic clustering on an obtained word after word segmentation and stop word removal are separately performed on at least one type of the following information;

posted information, favorited information, reposted information, and comment information of the target user, and determining the point of interest of the target user by using a clustering result.

14. The computer system according to claim 10, wherein the instructions further cause the processor to determine the quality user related to the point of interest.

15. The computer system according to claim 14, wherein the instructions further cause the processor to:

select a user whose score meets a preset requirement from a user who has an association relationship with a predetermined seed user related to the point of interest, as the quality user, wherein the score is determined according to one of or any combination of the following factors;

a degree of content matching between information that is posted, reposted, or self-described by the user and the point of interest, an activity degree of the user, impact of the user, and a spread range of user information.

16. The computer system according to claim 10, wherein the instructions further cause the processor to:

compute semantic similarity between the point of interest and each piece of information in the historical operation information of the quality user corresponding to the point of interest, and select information whose semantic similarity meets a preset similarity requirement as the information related to the point of interest.

17. The computer system according to claim 10, wherein when the interest classification model is trained, a used feature comprises a keyword of information.

18. The computer system according to claim 17, wherein when the interest classification model is trained, the used feature further comprises one of or any combination of the following factors:

an information poster, a person followed by the information poster or a follower of the information poster, a quantity of comments on information, a quantity of reposts of information, and a quantity of favorites of information.

19. The computer system according to claim 10, wherein the to-be-recommended information is real-time information, notification information, or promotion information.