An Improved Hybrid Approach for Proactive Healthcare Industry Risk Prediction: Integrating Graph Weighted Fusion, RF...

Area of Science:

• Healthcare logistics
• Supply chain management
• Network science

Background:

• Traditional healthcare supply chain risk prediction often neglects enterprise-level data and interdependencies.
• Existing methods lack the granularity for fine-grained enterprise risk assessment.
• Emerging 6G infrastructures necessitate real-time risk awareness and prediction capabilities.

Purpose of the Study:

• To propose IA-GWRG, a hybrid framework for fine-grained enterprise risk prediction in healthcare industrial chains.
• To integrate multi-relational graph modeling, key indicator selection, and subgraph-level learning.
• To leverage 6G capabilities for enhanced, real-time risk awareness.

Main Methods:

• Developed a salient graph feature extraction method using weighted fusion and a local-global-positional scheme.
• Implemented a key indicator identification module with a random forest algorithm for micro-level indicator selection.
• Employed a node risk assessment method combining indicators and structural embeddings via subgraph-level random walk and GCN-based learning.

Main Results:

• IA-GWRG demonstrated superior performance compared to state-of-the-art baselines on real-world healthcare datasets.
• The framework achieved high predictive accuracy and robustness.
• Effectiveness was validated for 6G-oriented deployment scenarios.

Conclusions:

• IA-GWRG offers a robust and effective solution for enterprise risk prediction in healthcare supply chains.
• The hybrid approach enhances interpretability and reduces feature noise.
• The framework is well-suited for real-time risk awareness in future 6G intelligent infrastructures.