Abstract
<jats:p>The use of particles modified epoxy is widely used in enhancing toughness and joint applications in engineering work. The different kinds of particles (rigid or soft, micro or nanoparticles) all have an effect to the overall mechanical properties in the failure situation. In order to help with interpretating the experimental results and provide more accurate predictions for engineering designs, machine learning framework with the use of Explainable AI (XAI) are used in this study. By using a Random Forest Regressor for mapping the relationships between the different design factors, such as particle types, weight percentage of particles use and failure modes against fracture energy outcomes, the patterns, trends and effects can be predicted for further design needs. The use of Explainable AI (XAI) via SHAP (SHapley Additive exPlanations) also helped to interpret if the model can be explained in real life situations, hence help to improve accuracy of results. The predictive model establishes strong alignment with lab experimental results, and SHAP insights provide true evaluations of the important features, model decision pathways, and distinct features for further studies.</jats:p>