Communication information processing method, system, device, medium and product
By combining multi-dimensional risk feature extraction and time-series analysis of large language models with deep Q-network optimization of response strategies, the problem of inaccurate risk analysis of communication information fraud was solved, and efficient hierarchical response to communication information was achieved, thereby reducing fraud risk.
CN122160093APending Publication Date: 2026-06-05CHINA MOBILE GROUP DESIGN INST +1
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA MOBILE GROUP DESIGN INST
- Filing Date
- 2026-01-26
- Publication Date
- 2026-06-05
AI Technical Summary
Technical Problem
Existing technologies are insufficient for accurate fraud risk analysis and response to communication information, making it difficult to effectively reduce communication fraud activities.
Method used
By acquiring the target user's communication information, multi-dimensional risk features are extracted and matched with pre-stored fraud features. Combined with a large language model, time-series risk analysis is performed to generate hierarchical response information. The response strategy is optimized using a deep Q-network to achieve hierarchical response processing of communication information.
Benefits of technology
It improves the accuracy of communication fraud risk analysis, reduces the risk of communication fraud, and enhances security.
✦ Generated by Eureka AI based on patent content.
Smart Images
Figure CN122160093A_ABST
Abstract
The embodiment of the application provides a communication information processing method, system, device, medium and product, and relates to the technical field of communication. The communication information processing method comprises the following steps: obtaining communication information of a target user; performing multi-dimensional risk feature extraction processing on the communication information to obtain multi-dimensional risk features; performing risk analysis processing on the communication information according to the multi-dimensional risk features and pre-stored fraud features to obtain a risk analysis result of the communication information; determining hierarchical response information for the communication information according to the risk analysis result; and performing hierarchical response processing on the communication information according to the hierarchical response information.
Need to check novelty before this filing date? Find Prior Art