Skip to main content

Table 4 Results obtained with the FS ensemble with subset size defined by RPT using β = 0.1 (upper values), β = 1 (middle values) and β = 10 (bottom values) for different classifiers. Results are averaged over 10 × 5 stratified cross validation, for each time window, using ADNI data

From: Neuropsychological predictors of conversion from mild cognitive impairment to Alzheimer’s disease: a feature selection ensemble combining stability and predictability

  

AUC

Sensitivity

Specificity

Stability

# Features

2Y

NB

0.758 ± 0.00

0.599 ± 0.01

0.834 ± 0.00

1.0 ± 0.0

2

0.839 ± 0.00

0.744 ± 0.01

0.779 ± 0.01

0.982 ± 0.01

17

0.864 ± 0.00

0.791 ± 0.01

0.819 ± 0.01

0.912 ± 0.01

35

SVM Poly

0.460 ± 0.08

0.323 ± 0.03

0.933 ± 0.00

1.0 ± 0.0

2

0.849 ± 0.00

0.758 ± 0.01

0.797 ± 0.01

0.982 ± 0.01

17

0.889 ± 0.01

0.789 ± 0.02

0.838 ± 0.01

0.913 ± 0.02

30

SVM RBF

0.770 ± 0.00

0.571 ± 0.02

0.829 ± 0.01

1.0 ± 0.0

2

0.864 ± 0.01

0.758 ± 0.02

0.825 ± 0.01

0.978 ± 0.02

22

0.891 ± 0.01

0.777 ± 0.02

0.841 ± 0.01

0.913 ± 0.02

30

DT

0.588 ± 0.02

0.446 ± 0.03

0.732 ± 0.03

1.0 ± 0.0

2

0.706 ± 0.02

0.574 ± 0.04

0.839 ± 0.01

0.934 ± 0.02

38

0.715 ± 0.03

0.568 ± 0.06

0.861 ± 0.01

0.919 ± 0.02

34

LR

0.769 ± 0.00

0.637 ± 0.01

0.784 ± 0.00

1.0 ± 0.0

2

0.846 ± 0.01

0.732 ± 0.03

0.821 ± 0.01

0.978 ± 0.02

22

0.882 ± 0.01

0.727 ± 0.02

0.848 ± 0.01

0.913 ± 0.02

32

3Y

NB

0.772 ± 0.01

0.521 ± 0.01

0.901 ± 0.00

1.0 ± 0.0

2

0.859 ± 0.00

0.761 ± 0.01

0.804 ± 0.01

0.985 ± 0.01

22

0.872 ± 0.00

0.775 ± 0.02

0.829 ± 0.01

0.889 ± 0.01

30

SVM Poly

0.734 ± 0.01

0.0 ± 0.0

1.0 ± 0.0

1.0 ± 0.0

2

0.879 ± 0.00

0.584 ± 0.02

0.925 ± 0.01

0.927 ± 0.01

37

0.876 ± 0.01

0.626 ± 0.02

0.912 ± 0.01

0.780 ± 0.02

55

SVM RBF

0.777 ± 0.01

0.169 ± 0.02

0.982 ± 0.01

1.0 ± 0.0

2

0.871 ± 0.01

0.614 ± 0.01

0.924 ± 0.01

0.985 ± 0.02

22

0.872 ± 0.01

0.619 ± 0.02

0.914 ± 0.01

0.942 ± 0.01

25

DT

0.602 ± 0.02

0.463 ± 0.02

0.742 ± 0.02

1.0 ± 0.0

2

0.704 ± 0.02

0.603 ± 0.03

0.804 ± 0.02

0.959 ± 0.03

22

0.719 ± 0.02

0.622 ± 0.03

0.816 ± 0.01

0.890 ± 0.02

33

LR

0.777 ± 0.01

0.505 ± 0.01

0.920 ± 0.00

1.0 ± 0.0

2

0.864 ± 0.01

0.658 ± 0.02

0.889 ± 0.01

0.985 ± 0.02

22

0.864 ± 0.01

0.658 ± 0.02

0.889 ± 0.01

0.985 ± 0.02

22

4Y

NB

0.858 ± 0.03

0.779 ± 0.01

0.820 ± 0.01

0.937 ± 0.02

15

0.891 ± 0.01

0.775 ± 0.01

0.844 ± 0.02

0.925 ± 0.02

36

0.886 ± 0.00

0.789 ± 0.01

0.819 ± 0.01

0.895 ± 0.02

32

SVM Poly

0.849 ± 0.01

0.706 ± 0.02

0.824 ± 0.02

0.937 ± 0.02

15

0.904 ± 0.00

0.757 ± 0.02

0.884 ± 0.01

0.846 ± 0.02

36

0.908 ± 0.01

0.757 ± 0.02

0.884 ± 0.01

0.847 ± 0.02

45

SVM RBF

0.871 ± 0.01

0.702 ± 0.01

0.887 ± 0.01

0.937 ± 0.02

15

0.901 ± 0.00

0.754 ± 0.02

0.863 ± 0.02

0.925 ± 0.02

10

0.905 ± 0.01

0.758 ± 0.01

0.873 ± 0.03

0.829 ± 0.03

70

DT

0.735 ± 0.03

0.708 ± 0.05

0.761 ± 0.02

0.937 ± 0.02

15

0.735 ± 0.03

0.708 ± 0.05

0.761 ± 0.02

0.937 ± 0.02

15

0.735 ± 0.03

0.708 ± 0.05

0.761 ± 0.02

0.937 ± 0.02

17

LR

0.870 ± 0.01

0.745 ± 0.02

0.862 ± 0.01

0.937 ± 0.02

15

0.870 ± 0.01

0.745 ± 0.02

0.862 ± 0.01

0.937 ± 0.02

15

0.869 ± 0.01

0.751 ± 0.02

0.846 ± 0.01

0.905 ± 0.02

25