Machine Learning‐Assisted Design of the Al–Co–Cr–Fe–Ni–Ti‐Based High‐Entropy Alloys With Target Room‐Temperature Tensile Elongation

The vast compositional space of Al–Co–Cr–Fe–Ni–Ti‐based high‐entropy alloys made it difficult to establish a one‐to‐one correspondence between alloy composition and mechanical properties using traditional trial‐and‐error methods. In the present study, a machine learning framework integrating feature engineering and data augmentation techniques has been established to thoroughly reveal the compositional dependence of room‐temperature tensile elongation in the Al–Co–Cr–Fe–Ni–Ti alloy system. By introducing empirical parameters and phase composition features and employing Pearson correlation analysis and feature selection along with exhaustive computation, the optimal feature subset comprising Al, Ti, VEC, δ r , and Δ H was determined. Moreover, the SMOTER algorithm was applied to balance the data distribution in order to address the sparsity of high‐elongation data, achieving a relatively high model coefficient of determination ( R 2 ≈ 90). Combined with SHAP and feature importance analyses, valence electron concentration (VEC) and atomic radius difference (δr) emerged as dominant factors in determining plasticity. The composition‐dependent machine learning model for tensile elongation proposed in this study demonstrated high prediction accuracy, reaching over 90%.