Construction of Gene Regulatory Networks Using Recurrent Neural Networks and Swarm Intelligence

Table 5

Results for the yeast dataset extracted from GNW with 50 time points and 25 time points represented as BAPSO_full and BAPSO_half.

	TP	TN	FP	FN	TPR	SPC	FPR	FNR	PPV	FDR	ACC	1-score	Graph edges	MSE	CPU time
BAPSO_full
0.5	6	75	13	6	0.50	0.85	0.15	0.50	0.32	0.68	0.81	0.39	19	0.0034	27.4 minutes
0.6	6	75	13	6	0.50	0.85	0.15	0.50	0.32	0.68	0.81	0.39	19
0.7	6	76	12	6	0.50	0.86	0.14	0.50	0.33	0.67	0.82	0.40	18
0.8	6	76	12	6	0.50	0.86	0.14	0.50	0.33	0.67	0.82	0.40	18
0.9	5	78	10	7	0.42	0.89	0.11	0.58	0.33	0.67	0.83	0.37	15
1.0	5	79	9	7	0.42	0.90	0.10	0.58	0.36	0.64	0.84	0.38	14
BAPSO_half
0.5	4	73	15	8	0.33	0.83	0.17	0.67	0.21	0.79	0.77	0.26	19	0.0048	14.7 minutes
0.6	4	73	15	8	0.33	0.83	0.17	0.67	0.21	0.79	0.77	0.26	19
0.7	4	74	14	8	0.33	0.84	0.16	0.67	0.22	0.78	0.78	0.27	18
0.8	4	74	14	8	0.33	0.84	0.16	0.67	0.22	0.78	0.78	0.27	18
0.9	4	76	12	8	0.33	0.86	0.14	0.67	0.25	0.75	0.80	0.29	16
1.0	4	77	11	8	0.33	0.88	0.13	0.67	0.27	0.73	0.81	0.30	15