Table 3 Top 10 important descriptors selected by four DT-based ensemble learning models
RF | ExtraTrees | ||
|---|---|---|---|
Selected descriptors | Feature importance | Selected descriptors | Feature importance |
Gap | 0.3412 | Gap | 0.0712 |
AP(xx) | 0.0986 | F01[C-N] | 0.0350 |
Chi1_EA(dm) | 0.0344 | AP(xx) | 0.0243 |
Chi0_EA(dm) | 0.0274 | SpMax2_Bh(i) | 0.0221 |
EP(xx) | 0.0239 | F02[C-N] | 0.0192 |
P_VSA_ppp_L | 0.0215 | SpMax7_Bh(m) | 0.0179 |
SpDiam_AEA(ed) | 0.0160 | F01[C-C] | 0.0174 |
SpMax_AEA(ed) | 0.0134 | C-004 | 0.0157 |
SpMin5_Bh(m) | 0.0119 | P_VSA_e_2 | 0.0152 |
CATS2D_06_LL | 0.0093 | EP(xx) | 0.0123 |
AdaBoost | GBM | ||
|---|---|---|---|
Selected descriptors | Feature importance | Selected descriptors | Feature importance |
Gap | 0.1196 | Gap | 0.1621 |
AP(xx) | 0.0682 | Solvent | 0.0534 |
P_VSA_MR_7 | 0.0601 | MATS1e | 0.0147 |
SpMax2_Bh(i) | 0.0408 | Chi1_EA(dm) | 0.0108 |
F01[C-N] | 0.0382 | MATS6m | 0.0107 |
P_VSA_ppp_L | 0.0317 | SpMax_AEA(ed) | 0.0102 |
F02[C-N] | 0.0224 | Eig01_AEA(ed) | 0.0100 |
LUMO | 0.0197 | CATS2D_00_LL | 0.0092 |
EP(xx) | 0.0167 | AP(xx) | 0.0089 |
SdsCH | 0.0141 | SpMin8_Bh(e) | 0.0086 |