AD ALTA
JOURNAL OF INTERDISCIPLINARY RESEARCH
Appendix 2. Classification performance indicators aggregated per feature selection and classification method, Source: author
feature selection
method
classification
method
train
ACC
(95 % CI)
train
AUC
(95 % CI)
train
TDL
(95 % CI)
test
ACC
(95 % CI)
test
AUC
(95 % CI)
test
TDL
(95 % CI)
CFS
LOGIT
0.872
(0.856,
0.888)
0.842
(0.802,
0.882)
4.013
(2.963,
5.063)
0.872
(0.853,
0.890)
0.837
(0.790,
0.884)
3.954
(2.753,
5.155)
CFS
CIT
0.930
(0.904,
0.956)
0.909
(0.857,
0.960)
5.964
(5.022,
6.907)
0.910
(0.890,
0.931)
0.865
(0.821,
0.910)
5.333
(4.477,
6.189)
CFS
SVM
0.924
(0.893,
0.956)
0.904
(0.870,
0.937)
5.874
(4.823,
6.926)
0.917
(0.892,
0.941)
0.894
(0.871,
0.917)
5.609
(4.817,
6.401)
CFS
RF
0.965
(0.927,
1.003)
0.990
(0.966,
1.014)
6.858
(6.215,
7.501)
0.922
(0.893,
0.951)
0.900
(0.878,
0.922)
5.788
(4.833,
6.743)
Consistency
LOGIT
0.890
(0.872,
0.907)
0.871
(0.848,
0.895)
4.827
(3.996,
5.658)
0.888
(0.869,
0.906)
0.868
(0.838,
0.898)
4.760
(3.908,
5.612)
Consistency
CIT
0.939
(0.924,
0.955)
0.925
(0.893,
0.956)
6.318
(5.800,
6.836)
0.922
(0.904,
0.939)
0.880
(0.847,
0.914)
5.759
(5.161,
6.358)
Consistency
SVM
0.943
(0.915,
0.971)
0.920
(0.894,
0.947)
6.457
(5.636,
7.277)
0.928
(0.906,
0.950)
0.900
(0.879,
0.922)
5.993
(5.282,
6.704)
Consistency
RF
0.984
(0.967,
1.001)
0.999
(0.994,
1.003)
7.057
(7.004,
7.109)
0.937
(0.919,
0.954)
0.910
(0.892,
0.928)
6.271
(5.689,
6.852)
FS
LOGIT
0.885
(0.863,
0.907)
0.856
(0.790,
0.923)
4.553
(3.682,
5.424)
0.884
(0.863,
0.904)
0.851
(0.765,
0.937)
4.507
(3.642,
5.372)
FS
CIT
0.922
(0.893,
0.951)
0.898
(0.828,
0.968)
5.673
(4.567,
6.780)
0.907
(0.878,
0.936)
0.862
(0.778,
0.946)
5.158
(4.041,
6.276)
FS
SVM
0.918
(0.881,
0.955)
0.891
(0.832,
0.950)
5.635
(4.364,
6.906)
0.910
(0.881,
0.938)
0.875
(0.802,
0.948)
5.331
(4.213,
6.449)
FS
RF
0.955
(0.908,
1.001)
0.982
(0.934,
1.031)
6.668
(5.707,
7.629)
0.916
(0.881,
0.950)
0.885
(0.814,
0.956)
5.496
(4.246,
6.746)
Relief
LOGIT
0.875
(0.856,
0.895)
0.825
(0.728,
0.923)
4.052
(3.072,
5.032)
0.875
(0.856,
0.893)
0.824
(0.737,
0.910)
4.016
(3.072,
4.961)
Relief
CIT
0.919
(0.887,
0.951)
0.889
(0.799,
0.978)
5.564
(4.343,
6.784)
0.902
(0.874,
0.930)
0.848
(0.763,
0.933)
4.955
(3.788,
6.122)
Relief
SVM
0.911
(0.876,
0.945)
0.881
(0.810,
0.951)
5.435
(4.245,
6.624)
0.903
(0.879,
0.927)
0.863
(0.789,
0.936)
5.143
(4.290,
5.996)
Relief
RF
0.955
(0.914,
0.997)
0.987
(0.967,
1.007)
6.726
(5.983,
7.468)
0.911
(0.879,
0.943)
0.875
(0.801,
0.949)
5.358
(4.215,
6.500)
IGR
LOGIT
0.896
(0.869,
0.923)
0.878
(0.824,
0.931)
5.000
(3.894,
6.106)
0.894
(0.868,
0.921)
0.875
(0.818,
0.932)
4.890
(3.713,
6.066)
IGR
CIT
0.937
(0.914,
0.960)
0.917
(0.875,
0.959)
6.225
(5.427,
7.024)
0.921
(0.898,
0.944)
0.880
(0.834,
0.926)
5.696
(4.910,
6.481)
IGR
SVM
0.938
(0.897,
0.979)
0.916
(0.861,
0.970)
6.305
(5.025,
7.585)
0.923
(0.893,
0.954)
0.896
(0.854,
0.939)
5.840
(4.816,
6.864)
IGR
RF
0.978
(0.943,
1.012)
0.996
(0.980,
1.012)
6.967
(6.460,
7.474)
0.934
(0.904,
0.965)
0.904
(0.862,
0.946)
6.183
(5.154,
7.213)
IG
LOGIT
0.898
(0.875,
0.920)
0.883
(0.852,
0.914)
5.064
(4.163,
5.965)
0.896
(0.874,
0.918)
0.879
(0.845,
0.913)
4.973
(4.051,
5.896)
IG
CIT
0.939
(0.920,
0.958)
0.921
(0.894,
0.948)
6.303
(5.692,
6.915)
0.921
(0.901,
0.941)
0.882
(0.846,
0.917)
5.717
(4.957,
6.477)
IG
SVM
0.941
(0.905,
0.978)
0.921
(0.885,
0.957)
6.411
(5.302,
7.519)
0.925
(0.900,
0.951)
0.900
(0.878,
0.922)
5.905
(5.026,
6.783)
IG
RF
0.980
(0.950,
0.997
(0.986,
7.005
(6.727,
0.936
(0.911,
0.908
(0.888,
6.249
(5.408,
- 65 -