Skip to main content

Table 3 Test errors for different classifiers

From: Addressing the Challenge of Defining Valid Proteomic Biomarkers and Classifiers

N train

SVM

Hierarchical Bayes

AdaBoost

Random Forest

Tree

GLM

7

44.7%

32.8%

43.2%

40.2%

48.1%

41.1%

20

40.2%

30.6%

40.2%

35.8%

43.2%

42.5%

33

32.0%

26.1%

23.1%

34.3%

38.1%

38.5%

67

16.4%

21.6%

14.1%

17.1%

26.0%

46.7%

  1. The test errors for different classifiers and different sizes of the training set (Ntrain). The results are based on a test set of 67 males and 67 females.