Skip to main content

Table 5 Median values (and standard deviation in parenthesis) of the performance metrics in test set by regularized LR methods across 100 runs applied to the full dataset

From: Identification of biomarkers predictive of metastasis development in early-stage colorectal cancer using network-based regularization

 

# Genes

Acc

Miscl/FN

Sensitivity

Specificity

AUC

# Common genes

EN

D1

59(32.63)

0.67(0.093)

6(1.667)/3(1.133)

0.67(0.126)

0.67(0.142)

0.67(0.093)

8

D2

45(21.34)

0.59(0.102)

7(1.732)/5(1.387)

0.38(0.173)

0.78(0.166)

0.58(0.095)

4

D3

39(19.76)

0.59(0.074)

7(1.257)/6(1.135)

0.25(0.142)

0.89(0.118)

0.57(0.072)

6

\(\bar{x}\)

48

0.62

0.43

0.78

0.61

6

iTwiner

D1

33(21.98)

0.78(0.075)

4(1.343)/4(1.362)

0.56(0.151)

1.00(0.036)

0.78(0.075)

19

D2

42(21.11)

0.65(0.056)

6(0.946)/6(0.904)

0.25(0.113)

1.00(0.040)

0.63(0.058)

25

D3

39(20.65)

0.65(0.050)

6(0.847)/6(0.783)

0.25(0.098)

1.00(0.040)

0.63(0.052)

30

\(\bar{x}\)

38

0.69

0.35

1

0.68

25

  1. D1—DATASET1; D2—DATASET2; D3—DATASET3; \(\bar{x}\)—datasets mean; # Genes—number of genes selected by the methods; Acc—accuracy; Miscl—misclassifications; FN—false negatives; Sensitivity—fraction of actual positive cases (P); Specificity—fraction of actual negative cases (PM); AUC—area under the ROC curve; # Common genes—number of genes selected in common in at least 50% of the runs