A method for complex network-oriented key node grouping discrimination and fusion sorting

By grouping node centrality indices of complex networks and constructing surrogate functions for filtering and merging sorting, the inconsistency and redundancy problems in existing technologies are solved, achieving stable and interpretable identification of key nodes, applicable to a variety of complex networks.

CN122241333APending Publication Date: 2026-06-19DALIAN UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
DALIAN UNIV
Filing Date
2026-01-27
Publication Date
2026-06-19

AI Technical Summary

Technical Problem

In existing technologies, key node identification methods for complex networks suffer from inconsistent evaluation, lack of a unified selection mechanism, and redundant indicators, resulting in unstable and difficult-to-interpret ranking results.

Method used

The node centrality index is divided into three groups: based on neighborhood information, path structure and iteration mechanism, a surrogate function is constructed for screening and fusion ranking, which comprehensively reflects the importance of nodes from multiple structural perspectives and suppresses index redundancy.

Benefits of technology

It achieves stable, reliable, and interpretable key node identification results, enhancing the interpretability and applicability of the method, and making it suitable for complex networks of different sizes and types.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122241333A_ABST
    Figure CN122241333A_ABST
Patent Text Reader

Abstract

This invention proposes a method for grouping and ranking key nodes in complex networks. First, it categorizes centrality indicators into three types based on their structural characteristics: centrality indicator groups based on neighborhood information, centrality indicator groups based on path structure, and centrality indicator groups based on an iterative mechanism. Within each category, a representative indicator selection process is performed. During this selection process, in addition to comparing the structural consistency based on surrogate quantities, the matching degree between each centrality indicator and its corresponding surrogate quantity in terms of distribution pattern, information entropy characteristics, and statistical consistency is comprehensively considered, thus achieving a balance between structural interpretability and statistical stability. Then, the ranking results of the representative centrality indicators selected from each category are fused, and a unified node criticality ranking result is generated through rank aggregation. This approach retains information from multiple structural perspectives while reducing the redundant influence between different centrality indicators.
Need to check novelty before this filing date? Find Prior Art