Skip to main content

Table 2 Performance after outlier detection in liver cancer mass spectrometry data.

From: A kernel-based approach for detecting outliers of high-dimensional biological data

Measurements Without outlier removal After outlier removal
   KLOD Mahalanobis One-class SVM
Specificity (%) 93.63 94.69 94.29 94.35
Sensitivity (%) 92.82 93.95 93.51 93.89
Accuracy (%) 93.14 94.23 93.82 94.07
No. of the outliers Cirrhosis 3 2 5
  HCC 2 4 6
\