| 赵凤媛,叶益聪,张周然,李亚豪,王洁,唐宇,李顺,白书欣.基于可解释机器学习模型的难熔高熵合金相预测[J].稀有金属材料与工程,2023,52(4):1192~1200.[Zhao Fengyuan,Ye Yicong,Zhang Zhouran,Li Yahao,Wang Jie,Tang Yu,Li Shun,Bai Shuxin.Interpretable Machine Learning Model-Based Phase Prediction for Refractory High-Entropy Alloys[J].Rare Metal Materials and Engineering,2023,52(4):1192~1200.] |
| 基于可解释机器学习模型的难熔高熵合金相预测 |
| 投稿时间：2022-09-20 修订日期：2023-03-23 |
| 中文关键词: 机器学习 难熔高熵合金 模型可解释性 固溶体 金属间化合物 |
| 基金项目:National Natural Science Foundation of China (Grant Nos. U20A20231 and 11972372) and College of Aerospace Science and Engineering Youth Talent Fund, National University of Defense Technology (Grant No. KY0505072209). |
| 中文摘要: |
| 采用k近邻 （KNN）、支持向量机（SVM）、决策树（DT）、随机森林（RF）和人工神经网络（ANN）5种机器学习（ML）方法对RHEAs中固溶体（SS）、混合固溶体和金属间化合物（SS+IM）进行了分类和预测。选择了5个输入相预测参数作为特征以及139组RHEAs数据以训练ML模型。结果表明，ANN模型的预测准确率最高，达到90.72%。9组新的四元和(TiVTa)xCr1–x体系RHEAs的实验结果显示，RF和ANN的预测精度更高，精准预测了11个SS和3个SS+IM合金的相组成。采用了SHAP（SHapley Additive exPlanations）模型来解释精度最高的 ANN 模型，并研究每个特征对相形成的贡献。5个特征的重要性顺序是混合焓（ΔHmix）、原子尺寸差（δ）、价电子浓度（VEC）、混合熵（ΔSmix）和电负性差（Δχ），其中ΔHmix的平均SHAP值大约是Δχ的5倍，是ΔSmix的4倍。较大的ΔHmix、较小的δ和VEC可能有助于RHEA中固溶体的形成。 |
| Interpretable Machine Learning Model-Based Phase Prediction for Refractory High-Entropy Alloys |
| 英文关键词:machine learning refractory high-entropy alloys model interpretability solid solution intermetallic |
| 英文摘要: |
| Five machine learning (ML) approaches, i.e. K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Decision Tree (DT), Random Forest (RF) and Artificial Neural Network (ANN) were used to classify and to predict the combination of phases, i.e. solid solutions (SS) and mixed solid solution and intermetallic (SS+IM) in refractory high-entropy alloys (RHEAs). Five input characteristic phase predicting parameters and 139 RHEAs were selected to train these models. Results show that ANN model has the highest accuracy of 90.72%. Experimental results of 9 quaternary and (TiVTa)xCr1–x RHEAs verify the accuracy of prediction and indicate that RF and ANN can predict more accurately, successfully predicting 11 SS and 3 SS+IM. SHAP (SHapley Additive exPlanations) model was used to interpret the ANN model which exhibits the highest accuracy and to investigate the contribution of each feature to phase formation. The order of importance of five features is enthalpy of mixing (ΔHmix), atomic size difference (δ), valence electron concentration (VEC), entropy of mixing (ΔSmix), and electronegativity difference (Δχ), where the mean SHAP value of ΔHmix is approximately 5 times higher than that of ?χ and 4 times higher than that of ΔSmix. Less negative ΔHmix, smaller δ and VEC may contribute to the formation of SS in RHEAs. |
|
|
| 摘要点击次数: 132 |
| 全文下载次数: 115 |
|
查看全文
查看/发表评论 HTML |
| 关闭 |
|
|
|
