Engineering the Energy Gap of Cupric Oxide Nanomaterial Using Extreme Learning Machine and Stepwise Regression Algorithms

Table 3

Performance of the developed intelligent and stepwise regression-based models and their performance superiority.

Developed model	Training phase			Testing phase
	CC	RMSE	MAE	CC	RMSE	MAE
ELM-IP-Sin	0.8302	0.4985	0.3312	0.8707	0.4761	0.3921
ELM-IP-Sig	0.8417	0.4828	0.3177	0.8382	0.5546	0.4321
SWR	0.1471	0.8844	0.7721	0.5930	0.9532	0.8840
% improvement of ELM-IP-Sig over ELM-IP-Sin	1.3661	3.1344	4.0578
% improvement of ELM-IP-Sin over ELM-IP-Sig				3.728839	14.1512	9.2570
% improvement of ELM-IP-Sin over SWR	82.28256	43.6399	57.1111	31.8939	50.0519	55.6512
% improvement of ELM-IP-Sig over SWR	82.52460	45.4064	58.8515	29.2559	41.8185	51.1270