TABLE 3.

Results of the prediction-based metrics with or without prior observation.

Models	Without prior observation				With one prior observation
	MDPE	MAPE	F20	F30	MDIPE	MAIPE	IF20	IF30
Serrano et al. (1999)	1.40	17.38	55.69	76.47	−2.62	9.46	71.70	90.57
Desoky et al. (2004)	11.25	24.28	43.53	57.25	0.47	10.64	69.81	83.02
Jiang et al. (2007)	66.84	66.84	12.94	20.39	10.98	12.07	66.04	81.13
Correa et al. (2008)	9.37	21.03	49.02	67.06	3.32	12.06	73.58	86.79
Williams et al. (2012)	−25.92	27.70	31.76	54.51	−19.61	19.61	52.83	77.36
Ogusu et al. (2014)	8.84	19.90	50.20	68.63	−2.23	11.11	71.70	84.91
Ding et al. (2015)	30.24	32.14	34.51	47.06	9.24	17.14	64.15	88.68
Rodrigues et al. (2018)	26.13	29.36	36.08	50.98	−5.02	13.39	71.70	100
Gu et al. (2021)	4.35	18.74	51.76	67.06	−7.61	14.56	62.26	79.25
Teixeira-da-Silva et al. (2022)	1.34	19.80	50.20	67.06	−14.20	15.83	56.60	81.13
Impact of protein binding modeling strategy
One-binding site model	3.48	19.38	51.37	68.24	−3.05	16.29	61.70	85.11
Langmuir equation	0.82	20.49	47.84	67.45	−1.73	4.06	97.87	100
Dose-dependent maximum effect model	−1.10	27.67	39.61	52.94	−4.43	5.47	97.87	97.87
Linear non-saturable binding equation	−0.19	19.8	50.20	72.16	−2.51	5.51	97.87	100
The simple exponent model	1.50	17.68	56.47	72.94	−2.79	6.47	97.87	97.87

F ₂₀ , F ₃₀ , the percentages of prediction error within 20% and 30%, respectively; MAPE, median absolute prediction error; MDPE, median prediction error; IF ₂₀ , IF ₃₀ , the percentages of individual prediction error within 20% and 30%, respectively; MAIPE, median absolute individual prediction error; MDIPE, median individual prediction error.