FRF-FHG: Constructing Fuzzy Hypergraphs Based on FRF

DOI: https://doi.org/10.62517/jike.202604130

Author(s)

Ziqian Zhao*, Xiaotong Zhang

Affiliation(s)

College of Science, North China University of Science and Technology, Tangshan, China *Corresponding Author

Abstract

The partitioning mechanisms of fuzzy hypergraphs predominantly rely on empirical rules, lacking explicit traceable paths that constrain model credibility and generalizability. To address this, we propose FRF-FHG, a fuzzy hypergraph construction method based on Fuzzy Random Forest (FRF). By building FRF from data and extracting complete branch paths from root to leaf nodes for each tree, we define fuzzy hyperedges as sample sets satisfying all path feature constraints, achieving effective mapping from rule space to hypergraph structure. To mitigate path redundancy, exponential growth of fuzzy hyperedges, and "densification" issues caused by FRF's inherent branching mechanism, we design a fuzzy hyperedge reduction strategy to precisely retain core association information between samples. Comparative and ablation experiments conducted on four cross-domain datasets demonstrate that FRF-FHG achieves robust and superior classification performance.

Keywords

FRF; Fuzzy Hypergraph; Fuzzy Hyperedge Reduction; Classification Model

References

[1] Zhang Z, Mei X, Jiang H, et al. Dynamical analysis of Hyper-SIR rumor spreading model[J]. Applied Mathematics and Computation, 2023;446:127887. [2] Fang Y, Zheng Y. Metric learning based on attribute hypergraph[C]//2017 IEEE International Conference on Image Processing (ICIP). IEEE, 2017;3440-3444. [3] Yin T, Ding W, Ju H, et al. The fuzzy hypergraph neural network model based on sparse k-nearest neighborhood granule[J]. Applied Soft Computing, 2025;112721. [4] Wei J X, Long C, Fu H, et al. Malicious Domain Name Detection Method Based on Enhanced Embedded Feature Hypergraph Learning[J]. Journal of Computer Research and Development, 2024;61(09):2334-2346. [5] Bianconi G, Dorogovtsev N S. Theory of percolation on hypergraphs[J]. Physical Review E, 2024;109(1-1):014306. [6] Shi J. A Research for Neighborhood Hypergraph Based Classification Algorithm for Incomplete Information System[D]. Chongqing University of Posts and Telecommunications, 2016. [7] Wang X, Yang Z, Han Y, et al. Method of sequential intention inference for a space target based on meta-fuzzy decision tree[J]. Advances in Space Research, 2024;74(8):4050-4067. [8] Qiu S, Wang Q, Xiong M, et al. Adaptive extraction of rotating Doppler shift based on random forest regression[J]. Physics Open, 2025;100347.