A Strategy for the Effective Optimization of Pharmaceutical Formulations Based on Parameter-Optimized Support Vector...

Area of Science:

• Pharmaceutical Sciences
• Computational Chemistry
• Drug Delivery Systems

Background:

• Pharmaceutical formulation development involves complex multiparametric optimization.
• Simplifying these optimization tasks is crucial for efficient drug development.
• Existing in silico methods require improvement for accuracy and reliability.

Purpose of the Study:

• To develop and validate a prediction model for optimizing pharmaceutical formulations.
• To compare the performance of the proposed model against other in silico methods.
• To demonstrate the model's utility in real-world formulation development.

Main Methods:

• Development of a prediction model using least square support vector machine (LSSVM) optimized by particle swarm optimization (PSO-LSSVM).
• Comparison with standard LSSVM and back propagation (BP) neural networks.
• Application of the PSO-LSSVM model to predict optimal formulations for quercetin solid dispersion and apigenin nanoparticles.
• Validation using Taguchi orthogonal design arrays for experimental optimization.

Main Results:

• The PSO-LSSVM model exhibited superior performance with the lowest mean square error on the test dataset.
• Predicted optimal formulations using PSO-LSSVM were consistent with experimental outcomes from Taguchi designs.
• The model demonstrated high reliability and usefulness in predicting formulation performance.

Conclusions:

• The PSO-LSSVM model offers a reliable and efficient approach for pharmaceutical formulation optimization.
• This predictive model can significantly reduce experimental workload and accelerate formulation design.
• The approach provides a cost-effective solution for optimizing drug preparation.