Skip to main content
. 2022 Jan 21;7(4):3713–3721. doi: 10.1021/acsomega.1c06389

Table 4. Descriptions of the 10 Data Sets Used in This Studya.

category data set # molecules # tasks task type
physical chemistry ESOL 1128 1 regression
  FreeSolv 642 1 regression
  Lipo 4200 1 regression
biophysics HIV 41 127 1 classification
  BACE 1513 1 classification
physiology BBBP 2039 1 classification
  Tox21 7831 12 classification
  ToxCast 8576 617 classification
  SIDER 1427 27 classification
  ClinTox 1478 2 classification
a

The columns “# molecules” and “# tasks” gave the numbers of molecule samples and the number of tasks of each data set. The column “task type” gave whether this data set was a regression or classification task. Further details on the data sets are available in Wu et al.19