# Table 3 Performance of prediction models

CCP
Class Sensitivity Specificity PPV NPV
HSC 0.92 0.943 0.939 0.926
LSC 0.943 0.92 0.926 0.939
BCC
Class Sensitivity Specificity PPV NPV
HSC 0.88 0.887 0.88 0.887
LSC 0.887 0.88 0.887 0.88
LDA
Class Sensitivity Specificity PPV NPV
HSC 0.9 0.943 0.938 0.909
LSC 0.943 0.9 0.909 0.938
NC
Class Sensitivity Specificity PPV NPV
HSC 0.9 0.962 0.957 0.911
LSC 0.962 0.9 0.911 0.957
SVM
Class Sensitivity Specificity PPV NPV
HSC 0.96 0.962 0.96 0.962
LSC 0.962 0.96 0.962 0.96
1. Sensitivity and specificity of compound covariate predictor (CCP), Bayesian compound covariate predictor (BCC), linear discriminator analysis (LDA), nearest centroid classification (NC), and support vector machines (SVM). Sensitivity is the probability for a class A sample to be correctly predicted as class A. Specificity is the probability for a non class A sample to be correctly predicted as non-A. Positive Predictive Value (PPV) is the probability that a sample predicted as class A actually belongs to class A. Negative Predictive Value (NPV) is the probability that a sample predicted as non class A actually does not belong to class A.
2. For some class high S100A8 cluster (HSC), if n11 = number of class HSC samples predicted as HSC, n12 = number of class HSC samples predicted as low S100A8 cluster (LSC), n21 = number of LSC samples predicted as HSC, and n22 = number of LSC samples predicted as LSC, then the following parameters can characterize performance of classifiers: Sensitivity = n11/(n11+n12), Specificity = n22/(n21+n22), PPV = n11/(n11+n21), and NPV = n22/(n12+n22). 