Table 13. Test set results of four base learners.

	Class	Precision	Recall	F1 score		Class	Precision	Recall	F1 score	Support
RF	1	0.8506	0.6720	0.7508	XGBoost	1	0.8632	0.9222	0.8917	3,085
	2	0.8707	0.9212	0.8952		2	0.9400	0.9153	0.9275	9,956
	3	0.8701	0.9106	0.8899		3	0.8705	0.9116	0.8906	9,956
	4	0.8424	0.8626	0.8524		4	0.8549	0.7901	0.8212	9,956
	5	0.7892	0.7118	0.7485		5	0.7284	0.7218	0.7251	5,108
	6	0.4164	0.2953	0.3455		6	0.1896	0.3512	0.2463	447
	Accuracy			0.8483		Accuracy			0.8503	38,508
	Macro avg	0.7732	0.7289	0.7471		Macro avg	0.7411	0.7687	0.7504	38,508
GDBT	1	0.8588	0.9151	0.8861	LightGBM	1	0.8577	0.9245	0.8899	3,085
	2	0.9314	0.9202	0.9258		2	0.9412	0.8962	0.9182	9,956
	3	0.8757	0.9011	0.8882		3	0.8697	0.8994	0.8843	9,956
	4	0.8147	0.8214	0.8180		4	0.8370	0.5416	0.6576	9,956
	5	0.8305	0.6658	0.7391		5	0.6730	0.7273	0.6991	5,108
	6	0.1539	0.3468	0.2132		6	0.0609	0.4698	0.1079	447
	Accuracy			0.8489		Accuracy			0.7803	38,508
	Macro avg	0.7442	0.7617	0.7451		Macro avg	0.7066	0.7431	0.6928	38,508

Note:

Values in bold represent the optimum values for each group.