Internet application account abnormal registration detection method and device and medium
By processing multi-dimensional correlated data through the number of SMS messages registered for internet application accounts and unsupervised outlier identification algorithms, abnormally registered internet accounts can be identified and processed. This solves the problem of operators being late in detecting and poorly identifying online fraud, and enables rapid and intelligent anomaly monitoring and handling.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA UNITED NETWORK COMM GRP CO LTD
- Filing Date
- 2026-04-01
- Publication Date
- 2026-06-16
AI Technical Summary
When faced with online fraud involving the registration and sale of internet accounts using mobile phone numbers, telecom operators suffer from problems such as late detection and poor identification. They need to accelerate the development of intelligent monitoring and identification technologies for users with abnormally registered internet accounts to improve their ability to handle the entire process, including prevention, early warning, and post-incident tracing.
By identifying suspected abnormal registration numbers based on the number of SMS messages used to register internet application accounts, an unsupervised outlier identification algorithm is used to process multidimensional correlated data, identify abnormal employee IDs, abnormal base stations, and abnormal terminals, and coordinate with internet application providers and operators for processing.
It has accelerated the detection speed of abnormal numbers, improved the effectiveness of abnormal handling, and achieved intelligent monitoring and accurate identification of abnormal registration behavior of Internet application accounts.
Smart Images

Figure CN122227242A_ABST