Predicting substrates of the human breast cancer resistance protein using a support vector machine method

BMC Bioinformatics

Table 4 Overlap of classification in 10 experimental models

Experimental model	98	87	77	63	91	45	42	62	7	73
98	100	86.31	87.07	86.69	84.41	81.37	87.07	87.07	87.45	86.69
87		100	87.83	85.93	87.45	81.37	87.83	87.07	87.45	90.49
77			100	89.73	94.30	87.45	91.64	92.34	89.73	91.26
63				100	87.83	84.79	88.21	89.73	89.35	88.59
91					100	88.59	88.97	94.30	91.64	92.40
45						100	85.93	88.21	84.79	84.79
42							100	91.64	92.78	92.02
62								100	91.26	91.26
7									100	91.26
73										100

The overall prediction accuracies (ACC) of the 10 experimental models 98, 87, 77, 63, 91, 45, 42, 62, 7, and 73 were 78.33%, 75.29%, 76.81%, 80.23%, 75.67%, 70.34%, 76.05%, 75.29%, 78.71%, and 77.95%, respectively.

ISSN: 1471-2105