. 2023 Jan 30;13:1666. doi: 10.1038/s41598-023-28770-2

Table 4.

Performance of the XGBoost model in comparison with the LSSVM model.

	Models	Train			Test			Overall
	Models	RMSE (cP)	R²	AARD (%)	RMSE (cP)	R²	AARD (%)	RMSE (cP)	R²	AARD (%)
Under saturated	XGBoost^a	0.040	0.999	0.576	0.048	0.999	1.194	0.042	0.999	0.699
Under saturated	LSSVM^b	0.030	0.999	1.500	0.040	0.999	1.400	0.040	0.999	1.400
Saturated	XGBoost^a	0.029	0.998	2.058	0.083	0.981	5.416	0.040	0.994	2.730
Saturated	LSSVM^b	0.310	0.988	13.500	0.770	0.838	13.200	0.380	0.979	13.480
Dead oil	XGBoost^a	0.632	0.928	7.018	0.748	0.867	12.542	0.431	0.931	7.982
Dead oil	LSSVM^b	1.780	0.959	21.300	1.650	0.914	19.700	1.820	0.955	21.200

^aXGBoost model (This Study).

^bLSSVM model¹⁹.