Table 2 Performance of the prediction of the ADME endpoints by ML model evaluated by the MSE and Pearson’s r value (MSE/r-value)
From: Understanding predictions of drug profiles using explainable machine learning models
Data set | RT | RF | NN | LightGBM |
|---|---|---|---|---|
HLM | 0.31/0.47 | 0.25/0.59 | 0.36/0.30 | 0.23/0.63 |
RLM | 0.44/0.52 | 0.35/0.63 | 0.52/0.32 | 0.33/0.66 |
hPPB | 0.38/0.58 | 0.27/0.74 | 0.49/0.32 | 0.25/0.76 |
rPPB | 0.38/0.57 | 0.30/0.67 | 0.55/0.21 | 0.28/0.70 |
MDR1 | 0.32/0.60 | 0.26/0.71 | 0.38/0.48 | 0.23/0.74 |
Sol | 0.50/0.35 | 0.40/0.54 | 0.55/0.16 | 0.37/0.59 |