Author(s)

Lei Zhang

Affiliation(s)

School of Traffic Management Engineering, Guangxi Police College, Nanning 530028, Guangxi, China.

Corresponding Author

Lei Zhang

ABSTRACT

The phase formation behavior of high-entropy alloys plays a decisive role in their mechanical, thermal, and chemical properties. This paper proposes an interpretable machine learning framework for predicting the phase formation of aluminum-containing HEAs. A dataset of 2256 alloy data points was constructed, covering three phase structures: body-centered cubic (BCC), face-centered cubic (FCC), and multiphase. Two gradient boosting models, Extreme Gradient Boosting and Light Gradient Boosting Machine, were used for multi-class prediction. The model performance was evaluated using learning curves, accuracy, and F1-score using a five-fold cross-validation approach. The results show that the two models have similar prediction accuracy, with Extreme Gradient Boosting performing better with small sample sizes, while Light Gradient Boosting Machine exhibits stronger convergence stability with large sample sizes. To improve model interpretability, the SHAP interpretability method was introduced to quantify the contributions of different descriptors. The results show that the BCC phase is primarily determined by the pure chemical enthalpy Hli and the valence electron concentration VEC; the FCC phase is significantly influenced by VEC, atomic number AN, melting point Tm, and ionization energy Eion; and the Multiphase is dominated by the mixing entropy Smix and the atomic number difference ΔAN. These findings not only verify existing empirical rules for phase selection in high-entropy alloys, but also reveal the role of aluminum in enhancing lattice distortion and entropy-driven stability. They also demonstrate the effectiveness of combining the interpretable gradient boosting model with SHAP, providing a reliable tool for the accelerated design of high-entropy alloys.

KEYWORDS

High-entropy alloys; Phase prediction; Machine learning; SHAP analysis

CITE THIS PAPER

Lei Zhang. Machine learning-enhanced prediction of phase formation in aluminum-containing high-entropy alloys: a comprehensive study with interpretable feature analysis. Journal of Computer Science and Electrical Engineering. 2025, 7(6): 66-70. DOI: https://doi.org/10.61784/jcsee3091.

REFERENCES

[1] Yeh J W, Chen S K, Lin S J, et al. Nanostructured high-entropy alloys with multiple principal elements: novel alloy design concepts and outcomes. Advanced Engineering Materials, 2004, 6(5): 299-303.

[2] Li J, Zhang W, Gao Z. Advances in Machine Learning-aided Design of High-entropy Alloys. Materials Reports, 2025, 39(18): 24080091-10.

[3] Yan Y, Hu X, Liao Y, et al. Recent machine learning-driven investigations into high entropy alloys: A comprehensive review. Journal of Alloys and Compounds, 2025, 1010: 177823.

[4] Yu B, Ren Y, Zeng Y, et al. Recent progress in high-entropy alloys: A focused review of preparation processes and properties. Journal of Materials Research and Technology, 2024, 29: 2689-2719.

[5] Zeng Y, Man M, Bai K, et al. Revealing high-fidelity phase selection rules for high entropy alloys: A combined CALPHAD and machine learning study. Materials & design, 2021, 202: 109532.

[6] Karathanasopoulos N, Singh A, Hadjidoukas P. Machine learning-based modelling, feature importance and Shapley additive explanations analysis of variable-stiffness composite beam structures. Structures, Elsevier, 2024, 62: 106206.

[7]  Zhang L, Huang J. Efficient Phase Classification in Aluminum-Containing High Entropy Alloys Using Machine Learning. 2024 IEEE First International Conference on Data Intelligence and Innovative Application (DIIA), Nanning, China. 2024, 1-6.

[8] Huang X, Jin C, Zhang C, et al. Machine learning assisted modelling and design of solid solution hardened high entropy alloys. Materials & Design, 2021, 211: 110177.

[9]  Wen C, Zhang Y, Wang C, et al. Machine learning assisted design of high entropy alloys with desired property. Acta Materialia, 2019, 170: 109-117.

[10]  Chen T, Guestrin C. Xgboost: A scalable tree boosting system. Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, 2016, 785-794.

[11]  Cortes C, Vapnik V. Support-vector networks. Machine learning, 1995, 20(3): 273-297.

[12]  Sun Y, Ni J. Machine learning advances in high-entropy alloys: a mini-review. Entropy, 2024, 26(12): 1119.

[13] Sokolova M, Lapalme G. A systematic analysis of performance measures for classification tasks. Information processing & management, 2009, 45(4): 427-437.

[14]  Lundberg S M, Lee S I. A unified approach to interpreting model predictions. Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS'17). Curran Associates Inc., Red Hook, NY, USA, 2017, 4768-4777.

[15]  Guo S, Ng C, Lu J, et al. Effect of valence electron concentration on stability of fcc or bcc phase in high entropy alloys. Journal of applied physics, 2011, 109(10): 103505.

[16]  Khoa N L, Khanh N D, Quyen H T N, et al. Thermodynamic and Electronic Descriptor-Driven Machine Learning for Phase Prediction in High-Entropy Alloys: Experimental Validation. Journal of Powder Materials, 2025, 32(3): 191-201.