Predicting drug solubility in supercritical carbon dioxide green solvent using machine learning models based on thermodynamic properties

Abstract

Reliable prediction of drug solubility in supercritical carbon dioxide (scCO₂) is crucial for the efficient design of pharmaceutical processes, including particle engineering and supercritical fluid-based extraction. Given that experimental determination of drug solubility in scCO₂ is costly and time-consuming, this study employs machine learning models to predict drug solubility in scCO₂, offering the advantage over thermodynamic models and empirical correlations of being able to predict the solubility of drugs beyond the model’s training range. In this work, authors use CatBoost, XGBoost, LightGBM, and RF models to predict the solubility of a set of drugs (Sixty-eight) in scCO2. Statistical errors and graphical analyses showed that the XGBoost model performed better than other models and had high reliability for predicting solubility. Among the evaluated models, XGBoost delivered the most accurate predictions, achieving a root mean square error (RMSE) of just 0.0605 and an R² value of 0.9984. Notably, 97.68% of the data points fell within the model’s applicability domain, highlighting its strong predictive reliability. These outcomes underscore the capability of the XGBoost algorithm to serve as a robust and efficient approach for estimating solubility.