Solubility prediction of lipid compounds using machine learning
Resumen:
Lipid purification processes are essential in lipid biomass valorization. Solubility is a key property in the solvent selection and process design. This work focuses on developing a predictive solubility model using machine learning techniques to optimize the separation of valuable compounds from a natural matrix derived from lanolin fat. First, the database was created from a literature review, then a database pre-processing step was performed, and the final step was model validation. Random Forest regression was selected for its ability to handle complex nonlinear relationships, showing better performance than bibliography models. An accurate model for lipids solubility in solvents was developed using machine learning techniques and experimental data.
| 2025 | |
|
CSIC Beca de Maestría ANII POS_NAC_2022_4_174069 |
|
|
Solubility Machine learning Data preprocessing Random Forest |
|
| Inglés | |
| Universidad de la República | |
| COLIBRI | |
| https://hdl.handle.net/20.500.12008/55260 | |
| Acceso abierto | |
| Licencia Creative Commons Atribución - No Comercial - Sin Derivadas (CC - By-NC-ND 4.0) |
| Sumario: | Lipid purification processes are essential in lipid biomass valorization. Solubility is a key property in the solvent selection and process design. This work focuses on developing a predictive solubility model using machine learning techniques to optimize the separation of valuable compounds from a natural matrix derived from lanolin fat. First, the database was created from a literature review, then a database pre-processing step was performed, and the final step was model validation. Random Forest regression was selected for its ability to handle complex nonlinear relationships, showing better performance than bibliography models. An accurate model for lipids solubility in solvents was developed using machine learning techniques and experimental data. |
|---|
