A Quantitative Prediction Model of Compounds on $\textbf{ER} \alpha$ Bioactivity Based on Regression Analysis

To construct the quantitative prediction model of compounds on
$ER_{\alpha}$
bioactivity, firstly, three mathematical re-gression analysis methods including stepwise regression, multiple linear regression, and multiple quadratic nonlinear regression are exploited; Secondly, we analyze the average error between the predicted and true values of the three regression prediction models on four training datasets; Finally, the least mean error regression model-stepwise regression-is used to predict the pIC50 values of 50 compounds in the test dataset.