Comparative Evaluation of Machine Learning-Assisted Statistical Modeling of Biopolymer Hydrogel Swelling for Material...

Area of Science:

• Materials Science
• Chemical Engineering
• Biotechnology

Background:

• Hydrogel swelling is critical for applications like drug delivery and wound dressings.
• Predictive modeling is essential for optimizing hydrogel formulation and performance.
• Sodium alginate hydrogels offer versatile properties for various technological uses.

Purpose of the Study:

• To investigate and optimize the swelling characteristics of sodium alginate biopolymeric hydrogels.
• To compare the predictive capabilities of Response Surface Methodology (RSM) and Artificial Neural Network (ANN) models.
• To establish a data-centric approach for intelligent hydrogel material design.

Main Methods:

• Experimental swelling degree measurements under controlled conditions (equilibrium and dynamic).
• Response Surface Methodology (RSM) with Central Composite Design (CCD) for initial analysis.
• Artificial Neural Network (ANN) model implementation for complex nonlinear associations.

Main Results:

• Both RSM and ANN models were evaluated using statistical parameters (R², RMSE).
• The ANN model demonstrated higher predictive accuracy and captured nonlinear trends better than RSM.
• RSM provided understandable equations and response surfaces for mechanistic insights.

Conclusions:

• Hybrid modeling approaches combining RSM and ANN offer complementary strengths for hydrogel studies.
• Machine learning and statistical design are synergistic for hydrogel system optimization.
• The study provides a methodological toolkit for hydrogel research and data-centric material design.