From: Prediction of probable backorder scenarios in the supply chain using Distributed Random Forest and Gradient Boosting Machine learning techniques
Performance metric
GBM (with actual data)
GBM (with ranged data)
DRF (with actual data)
DRF (with ranged data)
LogLoss
0.098
0.029
0.036
0.042
AUC
0.795
0.946
0.787
0.959
Mean per class error
0.423
0.07
0.430
0.103