	US 12,250,279 B2
	Method and apparatus for moderating abnormal users, electronic device, and storage medium
Yiyong Li, Guangzhou (CN); Qiushi Huang, Guangzhou (CN); Zhun Sun, Guangzhou (CN); Xue Jing, Guangzhou (CN); and Wei Xiang, Guangzhou (CN)
Assigned to BIGO TECHNOLOGY PTE. LTD., Singapore (SG)
Appl. No. 18/245,653
Filed by BIGO TECHNOLOGY PTE. LTD., Singapore (SG)
PCT Filed Aug. 30, 2021, PCT No. PCT/CN2021/115230 § 371(c)(1), (2) Date Mar. 16, 2023, PCT Pub. No. WO2022/068493, PCT Pub. Date Apr. 7, 2022.
Claims priority of application No. 202011066006.0 (CN), filed on Sep. 30, 2020.
Prior Publication US 2023/0336637 A1, Oct. 19, 2023
Int. Cl. H04L 67/50 (2022.01); G06F 17/18 (2006.01); H04L 67/1396 (2022.01)

CPC H04L 67/535 (2022.05) [G06F 17/18 (2013.01); H04L 67/1396 (2022.05)]

17 Claims

1. A method for moderating abnormal users, applicable to a device achieved by hardware or software and is integrated in an electronic device, the method comprising:

acquiring history behavior data of a plurality of to-be-moderated users, wherein the history behavior data comprises behavior data of interactions of the plurality of to-be-moderated users;

extracting a plurality of predetermined valid features from the history behavior data for each of the plurality of to-be-moderated users, wherein the predetermined valid features are features predetermined in sample data;

calculating, based on predetermined probabilities of events associated with the plurality of valid features, a probability that each of the plurality of to-be-moderated users is an abnormal user;

establishing a total probability function based on probabilities that the plurality of to-be-moderated users are the abnormal users;

determining candidate users by solving, based on predetermined conditions, a maximum value of the total probability function; and

determining the abnormal users by moderating the candidate users;

wherein determining the candidate users by solving, based on the predetermined conditions, the maximum value of the total probability function comprises:

initializing a set of abnormal users and a set of normal users, and organizing the plurality of to-be-moderated users to the set of abnormal users and the set of normal users, wherein a probability that each of the plurality of to-be-moderated users is the abnormal user or the normal user is acquired by inputting the history behavior data of the plurality of to-be-moderated users to a pretrained classification model, and each of the plurality of to-be moderated users is organized to the set of abnormal users or the set of normal users based on the probability that each of the plurality of to-be-moderated users is the abnormal user or the normal user;

calculating a first function value of the total probability function based on the set of abnormal users and the set of normal users;

traversing the set of abnormal users and the set of normal users, and recalculating a second function value of the total probability function;

updating the set of abnormal users and the set of normal users based on the first function value and the second function value, returning to processes of traversing the set of abnormal users and the set of normal users, and recalculating the second function value of the total probability function until the set of normal users is traversed, wherein a number of the to-be-moderated users in the set of abnormal users is less than a total number of the abnormal users in the plurality of to-be-moderated users and is greater than a product of the total number and a recall rate, the total number being a product of a predetermined proportion of the abnormal users and a number of the plurality of to-be-moderated users; and

determining the to-be-moderated users in the set of abnormal users as the candidate users;

wherein traversing the set of abnormal users and the set of normal users, and recalculating the second function value of the total probability function comprise:

traversing the set of normal users, and determining first to-be-moderated users currently traversed in the set of normal users as the abnormal users;

upon the traverse of the first to-be-moderated users in the set of normal users, traversing the set of abnormal users, and determining second to-be-moderated users currently traversed in the set of abnormal users as the normal users;

calculating, upon the traverse of the second to-be-moderated users in the set of abnormal users, the second function value of the total probability function based on the current set of abnormal users and the current set of normal users;

acquiring a plurality of second function values by repeating processes of traversing the set of abnormal users, determining the second to-be-moderated users currently traversed in the set of abnormal users as the normal users, and calculating, upon the traverse of the second to-be-moderated users in the set of abnormal users, the second function value of the total probability function based on the current set of abnormal users and the current set of normal users until the set of abnormal users is traversed; and

determining a minimum second function value in the plurality of second function values.