Skip to main content

Table 3 Classification performance of the three classification trees using different misclassification costs and the logistic regression models using different cut-off points

From: Use of data mining techniques to investigate disease risk classification as a proxy for compromised biosecurity of cattle herds in Wales

Classification trees Misclassification costs
  Tree 1:1 Tree 5:1 Tree 10:1 Tree 20:1
Sensitivity 0% 16% 64.3% 91%
PPV 0% 23.2% 15%. 11.9%
Specificity 100% 95% 65.5% 36.1%
NPV 91.3% 65.8% 95% 97.7%
Error rate 8.7% 11.9% 34.6% 59.1%
AUC   71% 71% 67.6%
Logistic regression Classification cut-offs
  0.5 0.3 0.2 0.1
Sensitivity 0% 1.1% 10.8% 65.5%
PPV 0% 22.3% 27.6% 15.5%
Specificity 100% 99.7% 97.4% 67.3%
NPV 91.64% 91.7% 92.3% 95.5%
Error rate 8.4% 8.6% 9.8% 32.8%
AUC 72% 72% 72% 72%