High-order dynamic Bayesian Network learning with hidden common causes for causal gene regulatory network

Lo, Leung-Yau; Wong, Man-Leung; Lee, Kin-Hong; Leung, Kwong-Sak

doi:10.1186/s12859-015-0823-6

BMC Bioinformatics

Table 11 Median delays F-scores of case III using long time series with the proposed algorithm with D-CLINDE

From: High-order dynamic Bayesian Network learning with hidden common causes for causal gene regulatory network

n	n _h	p _bias	T	em100	em200	em500	em1000	em2000	em5000
50	5	0.65	100	0.259	0.252	0.262	0.259	0.269	0.265
			200	0.430	0.420	0.431	0.435	0.426	0.431
			400	0.583	0.579	0.590	0.590	0.585	0.585
			800	0.753	0.758	0.752	0.757	0.759	0.750
			1000	0.774	0.787	0.781	0.777	0.772	0.780
			1200	0.801	0.791	0.799	0.806	0.811	0.805
			1400	0.815	0.824	0.823	0.822	0.825	0.822
			1600	0.831	0.843	0.837	0.839	0.833	0.835
		0.75	100	0.361	0.362	0.360	0.356	0.354	0.356
			200	0.477	0.484	0.485	0.488	0.486	0.494
			400	0.681	0.683	0.673	0.676	0.681	0.681
			800	0.789	0.803	0.787	0.782	0.795	0.785
			1000	0.809	0.818	0.820	0.818	0.818	0.821
			1200	0.821	0.816	0.830	0.828	0.820	0.831
			1400	0.827	0.835	0.830	0.831	0.834	0.832
			1600	0.824	0.830	0.835	0.828	0.829	0.828
		0.85	100	0.412	0.422	0.424	0.422	0.419	0.417
			200	0.569	0.565	0.555	0.554	0.555	0.573
			400	0.704	0.706	0.712	0.702	0.709	0.702
			800	0.806	0.805	0.803	0.803	0.801	0.807
			1000	0.794	0.795	0.798	0.796	0.789	0.795
			1200	0.818	0.820	0.819	0.825	0.822	0.820
			1400	0.824	0.822	0.822	0.813	0.819	0.822
			1600	0.821	0.827	0.813	0.818	0.826	0.821
100	10	0.65	1 00	0.291	0.277	0.282	0.290	0.283	0.285
			2 00	0.398	0.395	0.400	0.387	0.390	0.396
			4 00	0.566	0.575	0.576	0.571	0.571	0.574
			8 00	0.722	0.715	0.716	0.715	0.724	0.728
			1000	0.751	0.763	0.763	0.758	0.764	0.757
			1200	0.783	0.787	0.787	0.790	0.782	0.784
			1400	0.797	0.798	0.799	0.796	0.803	0.800
			1600	0.792	0.802	0.792	0.801	0.797	0.794
		0.75	100	0.360	0.370	0.358	0.362	0.363	0.356
			200	0.506	0.504	0.516	0.515	0.508	0.514
			400	0.688	0.690	0.689	0.692	0.689	0.700
			800	0.780	0.777	0.783	0.784	0.783	0.775
			1000	0.802	0.792	0.799	0.794	0.805	0.806
			1200	0.811	0.815	0.812	0.805	0.814	0.813
			1400	0.818	0.824	0.814	0.817	0.820	0.816
			1600	0.832	0.825	0.832	0.827	0.829	0.828
		0.85	100	0.412	0.426	0.424	0.419	0.415	0.408
			200	0.544	0.540	0.545	0.542	0.540	0.538
			400	0.695	0.689	0.690	0.690	0.691	0.692
			800	0.771	0.768	0.772	0.766	0.772	0.767
			1000	0.780	0.779	0.787	0.784	0.784	0.781
			1200	0.776	0.769	0.776	0.778	0.776	0.785
			1400	0.793	0.798	0.795	0.798	0.793	0.800
			1600	0.791	0.789	0.795	0.790	0.793	0.796

The results are on the incomplete data, with different number of iterations for the EM. em100 is using 100 EM iterations, em200 is using 200 EM iterations and so on.

Back to article page

ISSN: 1471-2105

Contact us

General enquiries: journalsubmissions@springernature.com