Skip to main content

Table 5 Classification results on benchmark datasets (± variance)

From: Correntropy induced loss based sparse robust graph regularized extreme learning machine for cancer classification

Datasets Evaluations Iris (L = 100) COIL20 (L = 500) USPST (L = 500) g50c (L = 1000) RNA-seq (L = 200)
RELM Acc 0.9511 ± 0.0021 0.9763 ± 0.0001 0.9229 ± 0.0000 0.9001 ± 0.0004 0.9275 ± 0.0021
Pre 0.9546 ± 0.0015 0.9753 ± 0.0004 0.9175 ± 0.0000 0.8937 ± 0.0014 0.9165 ± 0.0029
Recall 0.9495 ± 0.0021 0.9752 ± 0.0005 0.9140 ± 0.0000 0.9202 ± 0.0033 0.9279 ± 0.0012
F-mea 0.9459 ± 0.0025 0.9742 ± 0.0005 0.9150 ± 0.0000 0.9146 ± 0.0002 0.9205 ± 0.0020
L21-RFELM Acc 0.9622 ± 0.0006 0.9815 ± 0.0000 0.9337 ± 0.0002 0.8210 ± 0.0010 0.9525 ± 0.0003
Pre 0.9608 ± 0.0005 0.9823 ± 0.0000 0.9253 ± 0.0002 0.8407 ± 0.0034 0.9478 ± 0.0014
Recall 0.9650 ± 0.0004 0.9812 ± 0.0000 0.9251 ± 0.0002 0.8008 ± 0.0018 0.9531 ± 0.0002
F-mea 0.9604 ± 0.0005 0.9807 ± 0.0000 0.9240 ± 0.0002 0.8186 ± 0.0011 0.9497 ± 0.0002
LR21ELM Acc 0.9556 ± 0.0035 0.9814 ± 0.0000 0.9401 ± 0.0000 0.8150 ± 0.0023 0.9425 ± 0.0036
Pre 0.9594 ± 0.0019 0.9819 ± 0.0000 0.9341 ± 0.0001 0.8818 ± 0.0144 0.9445 ± 0.0029
Recall 0.9587 ± 0.0025 0.9803 ± 0.0000 0.9325 ± 0.0000 0.8040 ± 0.0164 0.9332 ± 0.0065
F-mea 0.9535 ± 0.0036 0.9799 ± 0.0000 0.9324 ± 0.0001 0.8423 ± 0.0057 0.9379 ± 0.0046
CELM Acc 0.9634 ± 0.0002 0.9865 ± 0.0000 0.9306 ± 0.0001 0.8674 ± 0.0002 0.9400 ± 0.0002
Pre 0.9702 ± 0.0002 0.9867 ± 0.0000 0.9215 ± 0.0001 0.8242 ± 0.0003 0.9458 ± 0.0001
Recall 0.9641 ± 0.0001 0.9851 ± 0.0000 0.9227 ± 0.0001 0.9344 ± 0.0002 0.9358 ± 0.0001
F-mea 0.9711 ± 0.0002 0.9853 ± 0.0000 0.9205 ± 0.0001 0.8758 ± 0.0002 0.9405 ± 0.0002
CSRGELM Acc 0.9788 ± 0.0000 0.9899 ± 0.0000 0.9513 ± 0.0005 0.9084 ± 0.0002 0.9625 ± 0.0002
Pre 0.9745 ± 0.0000 0.9897 ± 0.0000 0.9346 ± 0.0001 0.8964 ± 0.0003 0.9626 ± 0.0001
Recall 0.9747 ± 0.0000 0.9892 ± 0.0000 0.9325 ± 0.0001 0.9152 ± 0.0005 0.9631 ± 0.0002
F-mea 0.9735 ± 0.0000 0.9891 ± 0.0000 0.9325 ± 0.0001 0.9059 ± 0.0002 0.9622 ± 0.0001