Table 3.

The results of hyperparameters tunning in LGBC-based hybrid models development.

Prediction scenarios	Model	Hyperparameter
		$\mathrm{num}\mathrm{\_}\mathrm{leaves}$	$\max \mathrm{\_}\mathrm{depth}$	$\mathrm{learning}\mathrm{\_}\mathrm{rate}$	$n\mathrm{\_}\mathrm{estimators}$	$\max \mathrm{\_}\mathrm{bin}$
Scenario 1	LGAG	321	285	0.168	317	681,000
	LGBE	131	148	0.361	119	289,000
	LGGJ	792	45	0.747	796	844,000
	LGPO	613	999	0.504	999	1,000
Scenario 2	LGAG	654	213	0.566	813	685,100
	LGBE	999	648	0.736	889	753,000
	LGGJ	981	539	0.746	822	788,114
	LGPO	274	921	0.497	867	66,000
Scenario 3	LGAG	213	286	0.126	681	385,140
	LGBE	140	411	0.651	214	293,000
	LGGJ	442	845	0.740	226	922,419
	LGPO	299	633	0.771	77	253,000