Skip to main content

Table 3 Results of each algorithm model after PCA dimensionality reduction

From: Radiomics models for diagnosing microvascular invasion in hepatocellular carcinoma: which model is the best model?

  TP FN FP TN Accuracy Sensitivity Specificity
PCA + DT 24 7 9 18 72.41% 77.42% 66.67%
PCA + Bayes 12 11 9 26 65.52% 52.17% 74.29%
PCA + BPnet 16 10 5 27 74.14% 61.54% 84.38%
PCA + K-NN 17 10 7 24 70.69% 62.96% 77.42%
PCA + SVM 16 10 9 23 67.24% 61.54% 71.88%
PCA + RF 24 4 7 23 81.03% 85.71% 76.67%
PCA + GBDT 21 5 5 27 82.76% 80.77% 84.38%
  1. *FN False Negative, FP False Positive, TN True Negative, TP True Positive