Joint Decision-Making Model Based on Consensus Modeling Technology for the Prediction of Drug-Induced Liver Injury

Table 1

Parameters and performance of the 20 PaDEL submodels.

No.	Parameters		Performance (%)
					Training set				Fivefold cross-validation set				Test set
	C	g	ACC	SE	SP	MCC	ACC	SE	SP	MCC	ACC	SE	SP	MCC
1	2.10E + 06	11.0	96.4	96.4	96.5	92.4	70.5	72.2	67.6	38.8	74.1	75.3	71.9	46.2
2	6.45E + 07	4.3	97.8	98.6	96.4	95.2	71.3	73.0	68.3	40.3	71.2	72.6	68.9	40.5
3	6.00E + 06	14.4	95.5	95.7	95.2	90.4	70.0	69.9	70.0	38.7	74.4	77.4	69.4	46.2
4	6.58E + 07	41.9	92.1	92.8	91.0	83.2	70.6	71.9	68.5	39.4	72.0	75.0	66.8	41.2
5	4.87E + 07	10.4	97.7	97.9	97.4	95.1	69.9	69.9	69.8	38.6	74.2	74.4	74.0	47.1
6	9.46E + 07	34.8	94.4	94.4	94.4	88.1	71.2	69.2	74.6	42.4	71.8	72.6	70.4	41.9
7	3.25E + 07	22.3	96.3	96.0	96.6	92.0	69.6	69.0	70.7	38.5	72.0	73.2	69.9	42.1
8	8.80E + 07	32.3	94.0	94.3	93.6	87.3	70.0	70.7	68.8	38.4	72.2	74.7	67.9	41.8
9	1.67E + 07	17.3	94.5	94.2	94.9	88.3	70.5	69.4	72.3	40.4	72.3	72.3	72.4	43.5
10	8.97E + 07	29.6	95.6	95.7	95.6	90.7	70.0	69.6	70.5	38.9	73.1	74.1	71.4	44.5
11	2.18E + 07	13.3	97.5	98.0	96.7	94.7	70.0	70.0	69.8	38.7	72.5	73.8	70.4	43.2
12	2.39E + 07	23.8	94.9	95.0	94.7	89.1	70.9	70.2	72.2	41.1	71.2	72.0	69.9	40.8
13	6.63E + 07	27.7	95.0	95.0	95.1	89.5	70.1	70.3	69.8	38.9	72.5	73.8	70.4	43.2
14	6.33E + 07	36.5	93.7	93.9	93.2	86.5	70.1	69.5	71.1	39.4	72.0	72.9	70.4	42.2
15	9.85E + 07	41.0	94.2	94.1	94.3	87.7	70.1	69.9	70.6	39.3	72.3	72.6	71.9	43.3
16	3.88E + 07	24.1	95.0	95.1	94.7	89.3	71.0	69.7	73.1	41.4	73.7	74.7	71.9	45.6
17	6.27E + 07	28.3	95.2	95.6	94.7	89.8	70.3	69.6	71.7	39.9	72.2	73.5	69.9	42.4
18	4.33E + 07	26.9	95.9	96.4	94.9	91.2	70.3	70.5	70.0	39.3	73.7	73.8	73.5	46.0
19	2.47E + 07	26.5	94.0	94.5	93.2	87.3	70.5	69.5	72.3	40.5	73.7	73.5	74.0	46.2
20	6.56E + 07	37.1	92.7	93.3	91.7	84.4	69.7	69.6	69.7	38.2	73.9	73.2	75.0	46.8