Introduction: Liposomes are bilayered vesicles capable of encapsulating both hydrophilic and hydrophobic compounds, making them widely used in pharmaceuticals and cosmetics due to their excellent biocompatibility and versatility. However, they are structurally vulnerable to additives, such as ethanol and surfactants, which are often unavoidable during formulation. Therefore, it is essential to evaluate the effects of these components on liposomal stability and release behavior.
Method: Second-order multiple linear regression models were developed to predict liposomal release based on ethanol and five Tergitol™ 15-S surfactant concentrations. Nonlinear interactions were visualized using 3D regression surfaces. Liposomal stability was classified into four categories using K-nearest neighbors, logistic regression, and stochastic gradient descent algorithms. All models were implemented in Python using Scikit-Learn and Matplotlib.
Result: All regression models demonstrated high predictive accuracy, with R² values of 0.9611-0.9899 and mean absolute errors (MAE) of 2.19%-5.44%. No overfitting was observed. Among the classification models, logistic regression achieved the highest test accuracy (87.98%), followed by SGD (80.12%) and KNN (80.88%).
Discussion: Tergitol concentration had a greater impact on liposomal release than ethanol. Surfactants with higher HLB values showed weaker interactions with the lipid bilayer, resulting in reduced release. This aligns with previous findings that highly hydrophilic surfactants have limited bilayer penetration. The models effectively captured nonlinear interactions and offer practical utility for formulation prediction.
Conclusion: This study evaluated the stability of liposomes under various concentrations of ethanol and Tergitol surfactants and classified them using machine learning algorithms. The developed models can be effectively applied to formulation design in liposome-based systems, including pharmaceutical and cosmetic applications.
扫码关注我们
求助内容:
应助结果提醒方式:
