Skip to main content

Table 4 Simulation result for 100 nodes tree by comparing CBDN with other methods by AUC

From: Reconstructing directed gene regulatory network by only gene expression data

Covariance ARACNE CLR GENIE3 TIGRESS CBDN
(a) Simulation without  
any noise  
0.1 0.7445 0.8674 0.9388 0.8394 0.9804
0.2 0.9976 0.9995 1 0.8632 0.9231
0.4 1 1 1 0.8676 0.9792
0.6 1 1 1 0.8872 1
0.8 1 1 0.8426 0.9018 1
(b) Simulation with  
1/3 random noise  
0.1 0.6929 0.7572 0.8303 0.7765 0.8333
0.2 0.9561 0.9915 0.9992 0.8615 0.9894
0.4 1 1 1 0.8745 0.9875
0.6 1 1 1 0.9071 0.9905
0.8 1 0.9992 1 0.9511 0.9965
(c) Simulation with 2/3  
random noise  
0.1 0.4874 0.6362 0.6480 0.6547 0.9756
0.2 0.7527 0.8294 0.8867 0.8169 0.9794
0.4 0.9737 0.9871 0.9976 0.8843 0.9938
0.6 0.9990 0.9996 0.9998 0.9237 0.9907
0.8 0.9520 0.9973 0.9979 0.9123 0.9965