|
|
--- |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: train_data.csv |
|
|
- split: validation |
|
|
path: valid_data.csv |
|
|
- split: test |
|
|
path: test_data.csv |
|
|
license: cc-by-sa-4.0 |
|
|
--- |
|
|
|
|
|
Dataset Card for ESol (Estimated Solubility) Dataset |
|
|
|
|
|
Dataset Summary |
|
|
|
|
|
The ESOL dataset is designed for estimating the aqueous solubility |
|
|
of chemical compounds directly from their molecular structure. |
|
|
This dataset includes 2,874 experimentally measured solubility |
|
|
values. The most significant features for predicting solubility |
|
|
include calculated octanol-water partition coefficient (logP), |
|
|
molecular weight, the proportion of heavy atoms in aromatic systems, |
|
|
and the number of rotatable bonds. |
