fig1

Figure 1. Schematic workflow of the interpretable machine-learning-assisted composition-process co-design strategy for heat-resistant cast aluminum alloys: (A) data collection and database construction; (B) feature engineering and key-feature selection; (C) model training, testing, and performance evaluation; (D) genetic-algorithm-based optimization of alloy compositions and processing parameters; (E) SHAP-based model interpretation; and (F) experimental validation, including alloy preparation, heat treatment, elevated-temperature tensile testing, and microstructural characterization. SHAP: SHapley Additive exPlanations; UTS: ultimate tensile strength; R²: the coefficient of determination.