Prediction of pIC50 bioactive values based on knowledge priors and attention neural networks

Here's a refined version of your text: The ERα data set is utilized for predicting the bioactive values of pIC50, encompassing 1974 compounds. The dataset includes SMILES representations along with corresponding IC50 values or pIC50 values of ERα.

The IC50 (semi-inhibitory concentration) value serves as an indicator of the biological activity level, where lower values suggest greater effectiveness in inhibiting ERα. The pIC50 value is derived by taking the negative logarithm of IC50, with higher values indicating enhanced activity in inhibiting ERα. In this study, pIC50 is employed as the dependent variable for prediction.