摘要
使用基于支持向量机的机器学习方法,研究了金属中替代金属溶质与间隙轻溶质之间的结合能。元素的物理或化学属性被选择作为基本特征集。通过对特征进行优化,得到一组最优的特征组合。对数据集进行了拟合和留一交叉验证,并以极小的计算成本对数百个合金体系的结合能数据进行了预测。此外,对新体系的预测效果进行了测试。研究结果表明,基于支持向量机的机器学习方法能显著提高计算效率、加速材料科学研究。
In this paper,we study the binding energy between substitutional metal solutes and interstitial light solutes in metals using machine learning(ML)method based on support vector machine(SVM).The physical and chemical properties of the element are selected as the basic feature set.By optimizing input features,we get the optimal combination of features.Then,the fit and leave-one-out cross-validation are preformed.We predict the binding energies of hundreds of alloy systems with quite small computational cost.Furthermore,we measure the predictive effect for new systems.Our study demonstrates that the ML method based on SVM can significantly improve computational efficiency and accelerate materials science researches.
作者
何康尼
孔祥山
刘长松
HE Kang-ni;KONG Xiang-shan;LIU Chang-song(Key Laboratory of Materials Physics,Institute of Solid State Physics,Chinese Academy of Sciences,Hefei 230031,China;University of Science and Technology of China,Hefei 230026,China;School of Materials Science and Engineering,Shandong University,Jinan 250061,China)
出处
《安徽师范大学学报(自然科学版)》
CAS
2020年第5期409-414,429,共7页
Journal of Anhui Normal University(Natural Science)
基金
国家自然科学基金项目(11735015,51771185).
关键词
机器学习
支持向量机
溶质相互作用
结合能
machine learning
support vector machine
solute interaction
binding energy