Skip to main content

Table 2 Performance evaluations of models.

From: Computational models of liver fibrosis progression for hepatitis C virus chronic infection

Model Dataset CV test§ CA (%) Sensitivity (%) Specificity (%) F-measure (%)
LP Full (n = 42) 70/30-CV 93.20 82.00 100 90.11
LP-TOH TOH (n = 22) 70/30-CV 90.00 76.67 100 86.79
LP-IC IC (n = 20) 70/30-CV 95.00 85.00 100 92.31
  Validation test sets  
LP-TOH IC (n = 20) 90.00 71.43 100 83.33
LP-IC TOH (n = 22) 86.36 70.00 100 82.35
BNC-TOH IC (n = 20) 85.00 71.43 92.31 82.90
BNC-IC TOH (n = 22) 86.40 70.00 100 85.62
Rand BNC-TOH IC (n = 20) 45.00 na na na
Rand BNC-IC TOH (n = 22) 45.54 na na na
  1. LP model is based on the selected projection comprised of 9 HCV features. Classification accuracy for RandBNCs was averaged over 5 repetitions.
  2. §Values were averaged for 10 repetitions.