Table 3.

Prediction performance on GPCRs dataset for each pair of drug-target feature groups by the use of kNN classifier.

PCA	D i	Rec	Acc	Prec	F1
1	1	0.988	0.349	0.337	0.503
1	2	0.895	0.419	0.353	0.507
1	3	0.988	0.349	0.337	0.503
1	4	0.988	0.357	0.340	0.506
1	5	0.581	0.628	0.455	0.510
1	6	0.930	0.411	0.354	0.513
1	7	0.314	0.709	0.628	0.419
Average		0.812	0.460	0.401	0.494
2	1	1.000	0.353	0.340	0.507
2	2	0.895	0.422	0.355	0.508
2	3	0.988	0.345	0.336	0.501
2	4	0.988	0.349	0.337	0.503
2	5	0.570	0.636	0.462	0.510
2	6	0.907	0.426	0.358	0.513
2	7	0.279	0.717	0.686	0.397
Average		0.804	0.464	0.411	0.491
3	1	0.988	0.349	0.337	0.503
3	2	0.942	0.391	0.348	0.508
3	3	0.988	0.349	0.337	0.503
3	4	0.988	0.349	0.337	0.503
3	5	0.500	0.632	0.453	0.475
3	6	0.930	0.419	0.357	0.516
3	7	0.419	0.736	0.667	0.514
Average		0.822	0.461	0.405	0.503
4	1	0.988	0.349	0.337	0.503
4	2	0.895	0.415	0.352	0.505
4	3	1.000	0.345	0.337	0.504
4	4	1.000	0.349	0.339	0.506
4	5	0.558	0.632	0.457	0.503
4	6	0.942	0.419	0.358	0.519
4	7	0.314	0.713	0.643	0.422
Average		0.814	0.460	0.403	0.495
5	1	0.988	0.357	0.340	0.506
5	2	0.895	0.415	0.352	0.505
5	3	0.988	0.349	0.337	0.503
5	4	0.988	0.349	0.337	0.503
5	5	0.570	0.609	0.434	0.492
5	6	0.930	0.407	0.352	0.511
5	7	0.419	0.686	0.537	0.471
Average		0.825	0.453	0.384	0.499
6	1	0.988	0.349	0.337	0.503
6	2	0.895	0.426	0.356	0.510
6	3	0.988	0.349	0.337	0.503
6	4	1.000	0.345	0.337	0.504
6	5	0.535	0.640	0.465	0.497
6	6	0.942	0.419	0.358	0.519
6	7	0.337	0.733	0.707	0.457
Average		0.812	0.466	0.414	0.499
7	1	0.988	0.349	0.337	0.503
7	2	0.930	0.403	0.351	0.510
7	3	1.000	0.357	0.341	0.509
7	4	0.988	0.349	0.337	0.503
7	5	0.570	0.628	0.454	0.505
7	6	0.965	0.403	0.355	0.519
7	7	0.302	0.713	0.650	0.413
Average		0.820	0.457	0.283	0.495