Skip to main content

Table 3 Comparison of the confusion matrix of validated models

From: Modeling fuel break effectiveness in southern Spain wildfires

Observed

Predicted percentage correct

LR

DT

ANN

Non-effective fuel break

86.3%

83.6%

90.9%

Effective fuel break

88.3%

83.3%

87.5%

Overall percentage

87.2%

83.5%

89.3%

  1. LR logistic regression, DT decision tree, ANN artificial neural network