Table 6. The prediction results of candidate DNN classifiers generated by the solutions (pbest) and solutions of the DNN-PSO and DNN-NONPSO approaches, respectively, for one run of the 6–10 biological activity datasets.

	DNN-PSO					DNN-NONPSO
N of c	LOGD	METAB	NK1	OX1	OX2	LOGD	METAB	NK1	OX1	OX2
1	1.2482	1.2736	1.6554	1.0266	1.0777	1.2497	3.8073	0.9765	1.1638	1.9109
2	1.2408	1.2771	1.3910	1.0542	1.7795	1.2632	3.4689	0.4729	1.0406	1.9322
3	1.2406	1.2518	1.8322	1.0686	1.7323	1.2810	3.7857	1.7560	1.0377	1.8696
4	1.2675	0.4726	1.3428	1.0271	1.6467	1.2491	3.8208	0.9928	1.1308	1.0476
5	1.2496	0.4648	1.4885	1.5774	1.6804	1.2649	3.7294	0.5834	1.0956	1.0860
6	1.3532	1.2502	1.1518	1.0415	2.0970	1.2406	4.4178	1.4813	1.1960	1.1095
7	1.2599	0.4203	1.2815	1.2564	2.0083	1.2411	3.5244	2.1845	1.0281	1.6198
8	1.2498	0.4264	0.8911	1.1450	1.7527	1.2991	3.4270	1.4342	1.1475	2.0238
9	1.2447	1.2408	1.3539	1.1338	2.1255	1.3345	3.5422	1.6333	1.0672	2.0282
10	1.2424	0.7727	1.6205	1.0674	1.7241	1.2406	3.4181	0.5948	1.0979	1.9676
11	1.2470	0.4374	2.6143	1.0499	1.6180	1.2406	3.7073	1.5808	1.2062	1.7595
12	1.2463	1.2726	1.5508	1.2470	1.9858	1.3086	3.3049	2.1852	1.0790	1.0359
13	1.2604	1.2421	1.6819	1.2966	1.7688	1.2433	4.0172	1.6795	1.0230	1.8745
14	1.2424	1.2422	1.8027	1.0646	1.6546	0.4445	3.4941	0.7625	1.0265	1.8699
15	1.2441	1.2919	1.4408	1.0655	2.0290	1.2555	3.6298	1.1430	0.5670	1.8007
16	1.2475	1.2968	1.3721	1.0256	1.8031	1.3056	3.7405	1.7333	1.0231	1.8184
17	1.2788	1.2584	1.5148	1.0316	2.0049	1.2460	3.8127	1.4720	1.2042	1.5613
18	1.2440	1.2425	1.4503	1.0870	1.7276	1.2429	3.9439	1.5463	1.0480	1.8485
19	1.2465	0.4372	1.7415	0.5934	1.8387	1.2741	3.4115	0.5149	1.0456	1.8577
20	1.2448	1.2542	1.2924	1.0756	1.6917	1.2778	4.2795	1.6047	1.0486	1.8503
Avg	1.2549	0.9913	1.5235	1.0967	1.7873	1.2251	3.7142	1.3166	1.0638	1.6936
Std	0.0251	0.3841	0.3425	0.1787	0.2305	0.1858	0.2907	0.5305	0.1330	0.3391