Predicting Polymer Properties using Machine Learning: A Study on the Relationship between Molecular Structure and Mechanical Behavior

This article examines the application of machine learning (ML) methods to predict and Analysis of diverse physical properties of polymers using a rich dataset of polymers properties, this study covers a wide range of polymer properties, ranging from compressive and tensile strength to thermal and electrical behaviors. Using different regression methods Like Ensemble, Tree-based, Regularization and Distance-based, this research is done completely Evaluation using the most common quality criteria as a result of a series of empirical studies In choosing effective model parameters, those that provide a high-quality solution for it The stated problem was found. The best results were obtained by Random Forest with the highest R۲ scores of ۰.۷۱, ۰.۷۳ and ۰.۸۸ for glass transition, thermal decomposition and melting temperature, respectively. Results are intricately compared and provide valuable insights into performance Distinct ML approaches in predicting polymer properties predicted unknown values for each characteristic and method validation was performed by training the predicted values. Comparing the results with the specified variance values of each characteristic. Not only research It improves our understanding of polymer physics but also helps in informed model selection and optimization for materials science applications.

کلیدواژه ها:

Polymer Properties ، Machine Learning ، Molecular Structure ، Mechanical Behavior ، Regression Methods ، Random Forest ، Polymer Physics ، Materials Science