US 12,081,560 B2
Method and apparatus for preventing advertisement-related fraud, electronic device, and storage medium
Quan Cheng, Shenzhen (CN)
Assigned to TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Guangdong (CN)
Filed by TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED, Guangdong (CN)
Filed on Mar. 26, 2021, as Appl. No. 17/214,139.
Application 17/214,139 is a continuation of application No. PCT/CN2020/071413, filed on Jan. 10, 2020.
Claims priority of application No. 201910027586.3 (CN), filed on Jan. 11, 2019.
Prior Publication US 2021/0218761 A1, Jul. 15, 2021
Int. Cl. H04L 9/40 (2022.01); G06F 21/57 (2013.01); G06F 21/64 (2013.01)
CPC H04L 63/1416 (2013.01) [G06F 21/57 (2013.01); G06F 21/645 (2013.01); H04L 63/1483 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A method for preventing advertisement-related fraud that is performed by a server, the method comprising:
generating a user blacklist and an application (app) blacklist by applying a trained deep learning prediction model and model parameters obtained through training to historical advertisement response behavior of an advertising platform, the trained deep learning prediction model being configured to
(i) classify each user generating the historical advertisement response behavior as a normal user or a fraudulent user,
(ii) add each fraudulent user to the user blacklist, and
(iii) add an app to the app blacklist in response to a determination that a proportion of a quantity of clicks of all the fraudulent users on the app to a quantity of clicks of all advertisement clicking users and training sample users on the app is above a threshold;
obtaining behavior source information of a current advertisement response behavior, the behavior source information indicating a generation source of the current advertisement response behavior including a user identifier of a user that generated the current advertisement response and an app identifier of the app in which the user generated the current advertisement response;
determining whether the behavior source information of the current advertisement response behavior is fraudulent behavior source information based on a similarity between the user identifier and the app identifier of the behavior source information of the current advertisement response behavior and known fraudulent behavior source information including the user blacklist and the app blacklist; and
determining that the current advertisement response behavior is an advertisement-related fraudulent behavior when the behavior source information of the current advertisement response behavior is the fraudulent behavior source information.