Skip to main content

Table 4 Enrichment factor computed on the test set 2 (80 active molecules out of 3720 compounds)

From: Convolutional architectures for virtual screening

Architecture (Training 2) EF 1% EF 2
Tuned-MLP-Out 37(100%) 65(87.84%)
Voting 32(86.5%) 57(66.9%)
CNN 1D (F) 31(83.8%) 52(70.3%)
CNN 2D (T-L-E) 31(83.8% 52(70.3%)
RF(F) 37(100%) 62(83.8%)
SVM(F) 32(86.5%) 55(74.3%)
  1. Fingerprint types: (F) eatMorgan, (L) ayered, (T) orsion, (E)CFP4