ORIGINAL RESEARCH article
Front. Chem.
Sec. Chemical Physics and Physical Chemistry
Development of machine learning predictive models for estimating pharmaceutical solubility in supercritical CO2: Case study on Lornoxicam solubility
Provisionally accepted
- 1Inner Mongolia Minzu University, Tongliao, China
- 2Zhalantun vocational college, Zhalantun, China
- 3Tongling University, Tongling, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Production of nano-sized solid-dosage drugs is useful for pharmaceutical industry owing to high solubility and efficacy of the drugs for patients, which can also reduce the drugs side effects. For the solid-dosage oral formulations, the nanomedicine can be prepared via either top-down or bottom-up approach to enhance the drug solubility which in turns enhances the drug bioavailability. A novel methodology for simulation and prediction of medicine solubility in supercritical solvent was developed based on supervised learning algorithms for classification of the data. The data for the simulations were collected on solubility of a model drug in supercritical carbon dioxide. The supercritical-based processing is usually used for preparation of nanomedicine 2 with enhanced bioavailability, and the developed simulation method can help design and optimize the process for industrial applications. The data was obtained with temperature and pressure as the input parameters, whereas the drug solubility is considered as sole estimated output in the model. The validation outputs indicated that great agreement was obtained between the measured data and the simulated values with acceptable regression coefficient for the whole simulations. The simulation results revealed that the supervised learning algorithm is robust and rigorous for prediction of drug solubility data in supercritical conditions and can be used for process optimization and understanding the effects of process parameters. This study is innovative as it methodically assesses diverse machine learning methodologies, encompassing polynomial regression at different complexity tiers and the Gaussian Process Regressor for predicting pharmaceutical solubility. This comparative framework illustrates the bias-variance tradeoff and offers pragmatic guidance for choosing suitable models according to dataset attributes. The methodology presents a time-efficient and cost-effective alternative to conventional thermodynamic modelling for supercritical pharmaceutical processing.
Keywords: machine learning, pharmaceuticals, prediction, Nanotechnology, bioavailability
Received: 11 Aug 2025; Accepted: 27 Oct 2025.
Copyright: © 2025 meng, Wang, Erta, Guo, Wang, Zhao and Yang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Yuxia Yang, yuxia.yang584691@gmail.com
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.