A fuzzy node disambiguation method for academic networks

By constructing a fuzzy node disambiguation network, dynamically identifying and optimizing fuzzy nodes in academic networks, the problems of node ambiguity and insufficient information fusion are solved, the accuracy of node representation and the adaptability of the model are improved, and it can adapt to complex network structures.

CN119938999BActive Publication Date: 2025-09-30ANHUI UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202510033168.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-02-25
Publication Date
2025-09-30
Estimated Expiration
2045-02-25

AI Technical Summary

Technical Problem

Existing fuzzy node recognition methods in academic networks have problems such as limited node fuzziness disambiguation effect, insufficient adaptability in positive and negative sample selection, and insufficient fusion of local and global information, which affect the accuracy of node representation and the generalization ability of the model.

Method used

By introducing the attention mechanism, a fuzzy node disambiguation network is constructed, including node feature preprocessing and graph structure enhancement, fuzzy node recognition, contrastive learning and node representation optimization, and GNN node representation learning and classification modules. It dynamically identifies fuzzy nodes and optimizes node representation. It uses the self-attention mechanism and contrastive learning to adaptively select positive and negative samples, and fuse local and global information.

Benefits of technology

It significantly improves the accuracy of node classification and the robustness of the model, can adapt to complex and heterogeneous academic network structures, and enhances the discriminability of node representation and the generalization ability of the model.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The present invention discloses a fuzzy node disambiguation method for academic networks. The node feature preprocessing and graph structure enhancement module preprocesses the nodes of the academic network and constructs a graph structure. The graph structure is enhanced through an adjacency matrix. The enhanced graph is input into a fuzzy node identification module to dynamically generate attention weights. The attention weights are used to perform weighted aggregation on the features of neighboring nodes to generate a feature representation of the target node and identify the fuzzy nodes. All fuzzy nodes are input into a contrastive learning and node representation optimization module to eliminate the ambiguity of the fuzzy nodes and achieve node optimization. The optimized nodes are input into a GNN node representation learning and classification module for learning, and finally the node classification task is completed. By introducing an attention mechanism, the present invention solves the shortcomings of the existing technology in terms of adaptability in positive and negative sample selection and local and global information fusion.
Need to check novelty before this filing date? Find Prior Art