Jurisdictional enterprise network security management and control method and system and computer readable storage medium

CN121887542BActive Publication Date: 2026-05-29ZHEJIANG PONSHINE INFORMATION TECH CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
ZHEJIANG PONSHINE INFORMATION TECH CO LTD
Filing Date
2026-03-19
Publication Date
2026-05-29

AI Technical Summary

Technical Problem

Traditional cybersecurity management methods lack in-depth analysis of inter-enterprise relationships, are unable to identify complex hidden risks, have insufficient early warning capabilities, are passive in response, suffer from severe data silos, and are difficult to achieve intelligent and forward-looking security risk control.

Method used

By employing graph neural networks and large language models to fuse multi-source heterogeneous data from enterprises, an enterprise relationship graph is constructed. Dynamic risk assessment is performed through graph embedding vectors and risk scoring models, and interpretability technology is used to generate early warning information. In conjunction with privacy protection technology, cross-enterprise data collaborative analysis is conducted.

Benefits of technology

It enables in-depth insights into complex relationships between enterprises, enhances the cybersecurity protection capabilities of the jurisdiction, improves the accuracy of dynamic risk prediction and the interpretability of early warning, supports cross-enterprise collaborative protection, and protects data privacy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121887542B_ABST
    Figure CN121887542B_ABST
Patent Text Reader

Abstract

The present application relates to a kind of jurisdiction enterprise network security management and control method, system and computer readable storage medium, the multi-source heterogeneous network security data of each enterprise in jurisdiction is collected, and the enterprise correlation graph is constructed;Using graph neural network constructs and learns enterprise correlation graph, obtains the graph embedding vector of each enterprise;In the attention calculation of graph neural network, the real-time risk score of neighbor enterprise is as input;Using large language model, the semantic analysis of unstructured text data is carried out, the implicit correlation between enterprises and potential risk semantics are mined, and the enterprise correlation graph is updated based on implicit correlation and potential risk semantics;Based on the updated enterprise correlation graph, graph embedding vector and real-time security data, the dynamic risk score of each enterprise is calculated using fusion risk score model;Based on dynamic risk score, risk trend is predicted, and using explainable technology, early warning information containing risk cause explanation is generated.The present application significantly improves the network security protection capability of jurisdiction.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Knowledge graph-based industry chain power consumption analysis and prediction method, system and device, and medium

    CN121390438A

  • Enterprise risk assessment method and system based on large language model and high-frequency graph convolution

    CN121684635A