A complex network propagation threshold prediction method and system based on graph feature coding and statistical detection

By constructing a complex network structure and combining a graph feature encoding model with a statistical detection function, the problem of low efficiency in identifying propagation thresholds in complex networks in existing technologies is solved, and efficient and accurate propagation threshold prediction is achieved.

CN122432895APending Publication Date: 2026-07-21ZHEJIANG UNIV OF TECH
View PDF 4 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
ZHEJIANG UNIV OF TECH
Filing Date
2026-04-01
Publication Date
2026-07-21

AI Technical Summary

Technical Problem

Existing technologies suffer from high computational overhead and low efficiency when identifying critical parameters for propagation in complex networks, and they also struggle to accurately predict propagation thresholds when network structures are highly heterogeneous.

Method used

A method based on graph feature coding and statistical detection is adopted. By constructing a complex network structure, combining a graph feature coding model and an attention mechanism, propagation simulation and feature extraction are performed, and the propagation threshold is automatically identified using a statistical detection function.

Benefits of technology

It achieves accurate identification of critical parameters in the propagation process of complex networks while ensuring computational efficiency, reducing computational costs and improving the stability and applicability of identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122432895A_ABST
    Figure CN122432895A_ABST
Patent Text Reader

Abstract

A complex network propagation threshold prediction method and system based on graph feature coding and statistical detection, the method comprising: constructing a complex network structure for propagation simulation, initializing the number of network nodes and connection relationship; based on the susceptible-infected-susceptible propagation dynamics model, simulating the network propagation process under the condition of a preset propagation parameter set on the network structure, and introducing a fixed reference state recovery mechanism; when the propagation process reaches the preset time step, the node propagation state is intercepted and the corresponding next time propagation scale is counted, the mapping dataset between the propagation state and the propagation result is constructed and grouped; based on the propagation state sample, a propagation scale prediction model based on graph feature coding is constructed and trained, and the propagation scale prediction results under different propagation parameter conditions are obtained; statistical analysis is performed on the multiple propagation scale prediction results under the same propagation parameter condition, and a statistical detection function for describing the prediction output fluctuation characteristics is constructed; according to the response characteristics of the statistical detection function with the change of the propagation parameter, the position of the extreme value of the statistical detection function is determined, and the corresponding propagation parameter is taken as the propagation threshold prediction result of the complex network propagation process. The present application realizes the automatic identification of the critical parameter of the complex network propagation.
Need to check novelty before this filing date? Find Prior Art