Machine learning approaches to natural fiber composites: A review of methodologies and applications

Abstract

In recent years, the process of optimizing the design of natural fiber reinforcement in natural fiber composites (NFCs) with distinct properties has been redefined through the application of machine learning (ML). This work elucidates the functions of the types and applications of the ML algorithms and evolutionary computing techniques, with a particular focus on their applicability within the domain of NFCs. Moreover, the solution methodologies and associated databases were employed throughout various stages of the product development journey, from the raw material selection through the final end-use application for the NFCs. The strengths and limitations of the ML in the NFCs industry, together with relevant challenges, such as interpretability of ML models, in materials science was detailed. Finally, future directions and emerging trends in the ML are discussed.