Fig 1.
Corporate operational risk early warning indicator system.
Fig 2.
The construction process of the FCM clustering algorithm.
Fig 3.
Construction process of RF.
Table 1.
The parameter settings of the RF algorithm after parameter optimization.
Fig 4.
CRITIC weight calculation results for corporate operational risk warning indicators.
Fig 5.
Clustering evaluation results of the FCM algorithm.
Fig 6.
The membership distribution of various indicators in different risk categories by the FCM algorithm (Note: Clustering categories (Medium risk (2, 3)/ High risk (5)/ Risk-free (1)/ Low risk (4)).
Fig 7.
Corporate operational risk indicator statistical differences results.
Table 2.
Model performance comparison under different parameter settings.
Fig 8.
Parameter optimization results of the RF model improved by FCM clustering.
Fig 9.
Risk prediction results of the RF model improved by FCM clustering ((a) Different risk prediction results; (b) Prediction results of different models).
Fig 10.
Feature importance ranking results of model risk indicators.
Table 3.
The robustness test results of the model.