A hybrid higher-order graph convolutional network for fault diagnosis based on small sample label propagation

Key Points

• The proposed model enhances fault diagnosis accuracy through innovative label propagation methods with small sample sizes.
• Key evidence shows that the hybrid approach outperforms traditional models in terms of diagnostic performance metrics.
• Utilizing a higher-order graph convolutional network, this method integrates rich structural data to better identify faults.
• These findings highlight the potential for sophisticated model designs, especially where labeled data is scarce.