Stability of building gene regulatory networks with sparse autoregressive models

BMC Bioinformatics

Table 2 Performances of lasso, elastic-net (EN), and ridge MVAR models on synthetic datasets with varying numbers of genes and time points.

Number of genes	Number of time points	Method	True positives	False positives	Precision	Recall	F-measure	Stability
		EN	5.86	7.94	0.47	0.65	0.53	0.41
	10	Lasso	5.19	4.88	0.57	0.58	0.55	0.46
		Ridge	8.99	90.70	0.09	1.00	0.16	0.99
		EN	8.66	8.02	0.56	0.96	0.69	0.64
	30	Lasso	8.54	5.79	0.63	0.95	0.75	0.65
		Ridge	8.99	90.55	0.09	1.00	0.16	0.99
		EN	8.99	5.91	0.64	1.00	0.77	0.71
10	50	Lasso	8.99	4.12	0.71	1.00	0.82	0.75
		Ridge	8.98	90.53	0.09	1.00	0.16	0.99
		EN	9.00	5.12	0.67	1.00	0.79	0.72
	70	Lasso	8.99	3.92	0.72	1.00	0.83	0.76
		Ridge	8.99	90.35	0.09	1.00	0.16	0.99
		EN	17.29	131.50	0.12	0.35	0.18	0.15
	10	Lasso	13.53	42.81	0.25	0.28	0.26	0.15
		Ridge	48.90	2429.90	0.02	1.00	0.04	0.99
		EN	32.73	95.75	0.26	0.67	0.37	0.27
	30	Lasso	32.37	72.41	0.31	0.66	0.42	0.31
		Ridge	48.90	2431.70	0.02	1.00	0.04	0.99
		EN	39.39	77.88	0.34	0.80	0.48	0.35
50	50	Lasso	39.15	0.39	0.80	0.82	0.52	0.38
		Ridge	48.9	2436.20	0.02	1.00	0.04	0.99
		EN	48.82	98.11	0.34	1.00	0.50	0.41
	70	Lasso	48.82	78.38	0.39	1.00	0.56	0.45
		Ridge	49.00	2432.90	0.02	1.00	0.04	0.99
		EN	22.33	259.46	0.08	0.23	0.12	0.08
	10	Lasso	14.91	69.95	0.18	0.15	0.16	0.08
		Ridge	98.31	9759.90	0.01	1.00	0.02	0.98
		EN	58.67	214.17	0.22	0.59	0.32	0.20
	30	Lasso	56.37	153.76	0.27	0.57	0.37	0.22
		Ridge	98.71	9777.50	0.01	1.00	0.02	0.99
		EN	79.10	190.56	0.30	0.80	0.43	0.29
100	50	Lasso	78.60	164.55	0.33	0.79	0.46	0.32
		Ridge	98.80	9792.20	0.01	1.00	0.02	0.99
		EN	87.62	160.01	0.36	0.89	0.51	0.36
	70	Lasso	87.54	144.63	0.38	0.88	0.53	0.38
		Ridge	98.81	9805	0.01	1.00	0.02	0.99

ISSN: 1471-2105