A Social Media Account Risk Assessment System and Method Based on Friend Adding Behavior Simulation

Abstract

This invention provides a method for assessing the risk of a social media account, comprising: providing a simulator for simulating one or more behaviors of a virtual user on a social media platform; receiving an invitation from an unknown account through the simulator and presenting an account homepage of the unknown account based on the invitation; taking a screenshot of the account homepage through the simulator and extracting at least one text content and at least one image content from the screenshot; analyzing the at least one text content and at least one image content through a risk assessment module to determine at least one text feature and at least one image feature respectively; and determining, through the risk assessment module, whether the unknown account is low-risk or high-risk based on the at least one text feature and at least one image feature satisfying one of multiple conditions.

Claims

1. A method for assessing the risk of social media accounts based on simulating friend-adding behavior, comprising: A URL collector collects multiple URLs based on at least one filtering condition, each URL corresponding to a webpage containing a friend-adding link; an external API in a cloud environment provides these multiple URLs and their associated friend-adding links to an internal API in a local environment for storage; a simulator is provided to simulate one or more behaviors of a virtual user on a social platform; the internal API, in response to the multiple URLs and their associated friend-adding links, sends a request to a communication API of the social platform to assign the simulator to interact with an unfamiliar account associated with one of the friend-adding links; during the interaction, the simulator receives an invitation from an unfamiliar account and displays the unfamiliar account's homepage based on the invitation, the unfamiliar account originating from the friend-adding link; the simulator takes a screenshot of the account's homepage and extracts at least one text content and at least one image content from the screenshot. A risk assessment module analyzes at least one text content and at least one image content to determine at least one text feature and at least one image feature respectively; and the risk assessment module determines whether the unfamiliar account is low-risk or high-risk based on the at least one text feature and the at least one image feature satisfying one of multiple conditions.

2. The method as described in claim 1, wherein, The request instructs the emulator located on the local machine to provide a URL and the associated friend links to that URL.

3. The method as described in Request 2 further includes: receiving the URL and the friend links associated with the URL from the communication API of the social platform via the simulator, and interacting with the unknown account accordingly.

4. The method as described in claim 3 further includes: determining, based on an interaction result between the simulator and the unknown account, whether to transmit the at least one text content and the at least one image content extracted from the unknown account's homepage to a cloud database.

5. The method as described in claim 1, wherein, The at least one text feature and the at least one image feature include the text feature containing counterfeit keywords, the image feature containing different but similar corporate images, the account having insufficient friends, the unfamiliar account being an unofficial account, and the image feature containing celebrity images.

6. The method as described in Request 1 further comprises: providing the multiple URLs and their associated friend links to the internal API in the local environment via the external API located in the cloud environment in a text file, the text file containing an "urluid" for identifying each URL and the friend link associated with each URL.

Images