Skip to main content

Table 1 Performance of selected algorithms (%)

From: Classifying bridges for the risk of fire hazard via competitive machine learning

  AUC Training AUC Testing   No damage Damage Precision (%)
RF 73% 70% No damage 8/39/12 3/16/2 11/55/14
Damage 0/11/4 8/29/5 8/40/9
Recall (%) 8/50/16 11/45/7  
SVM 72% 68% No damage 7/37/4 4/18/10 11/55/14
Damage 1/11/0 7/29/9 8/40/9
Recall (%) 8/48/4 11/47/19  
GAM 71% 64% No damage 6/23/12 5/32/2 11/55/14
Damage 0/6/4 8/34/5 8/40/9
Recall (%) 6/29/16 13/66/7