Sample-Based Extreme Learning Machine with Missing Data

Table 6

Performance comparisons with UCI benchmark data sets. The two rows of each cell represent mean RMSE ± standard deviation over different missing ratios and training time (in seconds).


	S-ELM	ZF-ELM	MF-ELM

Body fat	0.1299 ± 0.0905	0.1781 ± 0.1007	0.1457 ± 0.1088
Body fat	0.1560	0.1758	0.1872

Housing	0.2229 ± 0.1019	0.3234 ± 0.1697	0.2511 ± 0.1261
Housing	0.5628	0.3432	0.3869

Pyrim	0.3816 ± 0.0842	0.5603 ± 0.2253	0.4537 ± 0.1396
Pyrim	0.0624	0.1008	0.0926

Abalone	0.2336 ± 0.1083	0.3004 ± 0.1537	0.2643 ± 0.1193
Abalone	213.5	177.6	182.4

Bike sharing	0.3886 ± 0.1020	0.4662 ± 0.1596	0.3977 ± 0.1218
Bike sharing	7.263	7.109	6.993

Airfoil self-noise	0.3765 ± 0.1591	0.4563 ± 0.2470	0.3810 ± 0.1928
Airfoil self-noise	52.38	50.74	51.61

Mathematical Problems in Engineering

Sample-Based Extreme Learning Machine with Missing Data

Table 6