BHO-XGBoost Prediction Model for Biological Activity of ERα Inhibitor Compounds

Breast cancer is the most common malignancy in women worldwide. The pathogenesis of the disease is closely related to the estrogen receptor alpha subtype (ERa). This paper presents a new optimization modeling method for breast cancer drug candidates. First of all, the research of random forest scoring mechanism was improved, using the maximum mutual information number - the Pearson correlation coefficient - distance correlation coefficient of three kinds of relevant indicators, select 20 each optimal molecular descriptors. Then, the Bayesian hyperparameter optimization method is used to optimize the parameters of extreme gradient boosting (XGBoost). The research results show that the model proposed in this paper has the best prediction performance compared with classical prediction models such as unoptimized XGBoost and SVR, and the goodness of fit is improved by at least 26.67%.