Skip to main content
. 2021 Jul 12;7:e577. doi: 10.7717/peerj-cs.577

Table 3. Detailed evaluation results.

In bold best accuracy values. CLFS and ELM metrics are similarity measures the biggest the best, MAE and EMD are distance measures the lowest the best.

Metric Size Type of source Event log GRU LSTM LSTM–GAN SIMOD
CLFS LARGE REAL POC 0.63141 0.67176 0.28998
BPI17W 0.63751 0.71798 0.36629 0.58861
BPI12W 0.58375 0.70228 0.35073 0.53744
CALL 0.82995 0.83043 0.24055 0.62911
SYNTHETIC CVS 0.83369 0.85752 0.20898 0.71359
CFM 0.81956 0.60224 0.11412 0.77094
SMALL REAL INS 0.50365 0.51299 0.25619 0.61034
ACR 0.78413 0.78879 0.18073 0.67959
MP 0.27094 0.23197 0.06691 0.34596
SYNTHETIC CFS 0.69543 0.66782 0.10157 0.76648
P2P 0.41179 0.65904 0.13556 0.45297
MAE LARGE REAL POC 801147 778608 603105
BPI17W 868766 603688 828165 961727
BPI12W 701892 327350 653656 662333
CALL 160485 174343 159424 679847
SYNTHETIC CVS 859926 667715 952004 1067258
CFM 25346 15078 956289 252458
SMALL REAL INS 1586323 1516368 1302337 1090179
ACR 344811 341694 296094 230363
MP 335553 321147 210714 298641
SYNTHETIC CFS 30327 33016 717266 15297
P2P 2407551 2495593 2347070 1892415
ELS LARGE REAL POC 0.58215 0.65961 0.28503
BPI17W 0.63643 0.70317 0.35282 0.58412
BPI12W 0.57862 0.67751 0.33649 0.52555
CALL 0.79336 0.81645 0.19123 0.59371
SYNTHETIC CVS 0.65160 0.70355 0.16854 0.70154
CFM 0.68292 0.43825 0.09505 0.66301
SMALL REAL INS 0.49625 0.50939 0.23070 0.57017
ACR 0.75635 0.45737 0.15884 0.71977
MP 0.25019 0.21508 0.04570 0.31024
SYNTHETIC CFS 0.54433 0.57392 0.07930 0.67526
P2P 0.22923 0.39249 0.09968 0.43202
EMD LARGE REAL POC 0.00036 0.00011 0.00001
BPI17W 0.00060 0.01010 0.00072 0.00057
BPI12W 0.00077 0.00061 0.00006 0.00002
CALL 0.00084 0.15794 0.00090 0.00072
SYNTHETIC CVS 0.61521 0.57217 0.40006 0.13509
CFM 0.00472 0.00828 0.03529 0.06848
SMALL REAL INS 0.03343 0.00308 0.33336 0.00001
ACR 0.49996 0.68837 0.25012 0.58674
MP 0.12609 0.33375 0.28577 0.31411
SYNTHETIC CFS 0.08253 0.10784 0.06924 0.03461
P2P 0.25306 0.33747 0.23898 0.03888