1
|
Original SBC implementation in the Verhaak samples present in the TCGA-GBM data set
|
K-means
|
None
|
80
|
80
|
4
|
6
|
6.406e-14
|
1.868e-02
|
0.573
|
0.744
|
2
|
Original SBC implementation in the TCGA-GBM dataset
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
4
|
3.197e-06
|
1.704e-03
|
0.478
|
0.683
|
3
|
ssGSEA on KEGG pathways for feature engineering
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
3
|
2.205e-01
|
6.616e-02
|
0.509
|
0.469
|
4
|
ssGSEA Oncogenic gene sets for feature engineering
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
4
|
3.958e-04
|
2.667e-02
|
0.501
|
0.634
|
5
|
ssGSEA on the Canonical Pathways for feature engineering
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
2
|
6.808e-03
|
1.157e-01
|
0.524
|
0.604
|
6
|
Penalised Cox model on the Oncogenic gene sets for feature engineering
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
5
|
3
|
0.000e+00
|
7.452e-02
|
0.530
|
0.966
|
7
|
PAFT the Oncogenic gene sets for feature engineering
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
4
|
0.000e+00
|
1.806e-02
|
0.528
|
0.959
|
8
|
Block HSIC-Lasso for feature selection
|
K-means
|
Karnofsky Index correction
|
160
|
261
|
4
|
4
|
1.078e-05
|
4.701e-02
|
0.565
|
0.698
|