Predictive Modelling of Polymer Properties from Polymer Genomic Data using Machine Learning Approaches

Polymer informatics, a burgeoning field in materials science, utilizes Machine Learning (ML) to predict properties based on complex associations between polymer structure and characteristics. ML, known for its adeptness with structured data, offers scalability and versatility across various polymers. This predictive approach holds transformative potential, streamlining materials discovery across industries. Despite success, challenges persist, including the need for diverse datasets, model interpretability, and generalization to new polymers. Addressing these challenges enhances the reliability of ML-based polymer property predictions. This research reviews recent advancements in ML-based polymer property prediction, detailing processes and techniques. Notably, it explores Polymer Genome, a web-based tool integrating ML algorithms for property prediction, fostering data accumulation, learning algorithms, and user interactions. The paper aims to deepen understanding of polymer behavior and the application of ML in predicting properties, contributing to the evolving landscape of polymer informatics.