Fig 1.
The flowchart of DeepHE.
Table 1.
Parameters of DeepHE.
Fig 2.
The distribution of essential genes across the 16 datasets.
Fig 3.
Performance comparison of DeepHE with different number of hidden layers and different dropout probability (DP) for two metrics: AUC and AP.
HL3 = [128, 256, 512], HL4 = [128, 256, 512, 1024], HL5 = [128, 256, 512, 1024, 1024]. DP: dropout probability.
Fig 4.
The ROC and PR curves of DeepHE with HL3 and DP = 0.2.
Fig 5.
Performance comparison of DeepHE with different class weights for two metrics: AUC and AP.
CW: class weight.
Fig 6.
Performance comparison of DeepHE with different features.
N+S: network embedding features plus sequence features; N: network embedding features; S: sequence features.
Fig 7.
Performance comparison of DeepHE, DC, BC, EC, and CC.
Fig 8.
Performance comparison of DeepHE with different feature combinations.
Fig 9.
Performance comparison of DeepHE, SVM, NB, RF, and Adaboost with different features.
N: network embedding features; S: sequence features.