基于可解释机器学习加速密度泛函理论的双金属过渡金属表面原子吸附能预测。

IF 2.5 4区 化学 Q2 CHEMISTRY, MULTIDISCIPLINARY
Jan Goran T. Tomacruz, Michael T. Castro, Miguel Francisco M. Remolona, Allan Abraham B. Padama, Joey D. Ocon
{"title":"基于可解释机器学习加速密度泛函理论的双金属过渡金属表面原子吸附能预测。","authors":"Jan Goran T. Tomacruz,&nbsp;Michael T. Castro,&nbsp;Miguel Francisco M. Remolona,&nbsp;Allan Abraham B. Padama,&nbsp;Joey D. Ocon","doi":"10.1002/open.202400124","DOIUrl":null,"url":null,"abstract":"<p>In this study, we identified features with the largest contributions and property trends in predicting the adsorption energies of carbon, hydrogen, and oxygen adsorbates on transition metal (TM) surfaces by performing Density Functional Theory (DFT)-based calculations and Machine Learning (ML) regression models. From 26 monometallic and 400 bimetallic fcc(111) TM surfaces obtained from Catalysis-hub.org, three datasets consisting of fourteen elemental, electronic, and structural properties were generated using DFT calculations, site calculations, and online databases. The number of features was reduced using feature selection and then finely-tuned random forest regression (RFR), gaussian process regression (GPR), and artificial neural network (ANN) algorithms were implemented for adsorption energy prediction. Finally, model-agnostic interpretation methods such as permutation feature importance (PFI) and shapely additive explanations (SHAP) provided rankings of feature contributions and directional trends. For all datasets, RFR and GPR demonstrated the highest prediction accuracies. In addition, interpretation methods demonstrated that the largest contributing features and directional trends in the regression models were consistent with structure-property-performance relationships of TMs like the d-band model, the Friedel model, and higher-fold adsorption sites. Overall, this interpretable ML–DFT approach can be applied to TMs and their derivatives for atomic adsorption energy prediction and model explainability.</p>","PeriodicalId":9831,"journal":{"name":"ChemistryOpen","volume":"14 4","pages":""},"PeriodicalIF":2.5000,"publicationDate":"2025-01-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/open.202400124","citationCount":"0","resultStr":"{\"title\":\"Atomic Adsorption Energies Prediction on Bimetallic Transition Metal Surfaces Using an Interpretable Machine Learning-Accelerated Density Functional Theory Approach\",\"authors\":\"Jan Goran T. Tomacruz,&nbsp;Michael T. Castro,&nbsp;Miguel Francisco M. Remolona,&nbsp;Allan Abraham B. Padama,&nbsp;Joey D. Ocon\",\"doi\":\"10.1002/open.202400124\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>In this study, we identified features with the largest contributions and property trends in predicting the adsorption energies of carbon, hydrogen, and oxygen adsorbates on transition metal (TM) surfaces by performing Density Functional Theory (DFT)-based calculations and Machine Learning (ML) regression models. From 26 monometallic and 400 bimetallic fcc(111) TM surfaces obtained from Catalysis-hub.org, three datasets consisting of fourteen elemental, electronic, and structural properties were generated using DFT calculations, site calculations, and online databases. The number of features was reduced using feature selection and then finely-tuned random forest regression (RFR), gaussian process regression (GPR), and artificial neural network (ANN) algorithms were implemented for adsorption energy prediction. Finally, model-agnostic interpretation methods such as permutation feature importance (PFI) and shapely additive explanations (SHAP) provided rankings of feature contributions and directional trends. For all datasets, RFR and GPR demonstrated the highest prediction accuracies. In addition, interpretation methods demonstrated that the largest contributing features and directional trends in the regression models were consistent with structure-property-performance relationships of TMs like the d-band model, the Friedel model, and higher-fold adsorption sites. Overall, this interpretable ML–DFT approach can be applied to TMs and their derivatives for atomic adsorption energy prediction and model explainability.</p>\",\"PeriodicalId\":9831,\"journal\":{\"name\":\"ChemistryOpen\",\"volume\":\"14 4\",\"pages\":\"\"},\"PeriodicalIF\":2.5000,\"publicationDate\":\"2025-01-29\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1002/open.202400124\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ChemistryOpen\",\"FirstCategoryId\":\"92\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/open.202400124\",\"RegionNum\":4,\"RegionCategory\":\"化学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q2\",\"JCRName\":\"CHEMISTRY, MULTIDISCIPLINARY\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ChemistryOpen","FirstCategoryId":"92","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/open.202400124","RegionNum":4,"RegionCategory":"化学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q2","JCRName":"CHEMISTRY, MULTIDISCIPLINARY","Score":null,"Total":0}
引用次数: 0

摘要

在这项研究中,我们通过基于密度泛函理论(DFT)的计算和机器学习(ML)回归模型,确定了在预测过渡金属(TM)表面上碳、氢和氧吸附能方面贡献最大的特征和性质趋势。从Catalysis-hub.org上获得的26个单金属和400个双金属fcc(111) TM表面中,使用DFT计算、现场计算和在线数据库生成了包含14个元素、电子和结构性质的三个数据集。利用特征选择减少特征数量,然后采用微调随机森林回归(RFR)、高斯过程回归(GPR)和人工神经网络(ANN)算法进行吸附能预测。最后,与模型无关的解释方法,如排列特征重要性(PFI)和形状加性解释(SHAP),提供了特征贡献和方向趋势的排名。对于所有数据集,RFR和GPR显示出最高的预测精度。此外,解释方法表明,回归模型的最大贡献特征和方向趋势与d波段模型、Friedel模型和高倍吸附位点等TMs的结构-性能-性能关系一致。总的来说,这种可解释的ML-DFT方法可以应用于TMs及其衍生物的原子吸附能预测和模型可解释性。
本文章由计算机程序翻译,如有差异,请以英文原文为准。

Atomic Adsorption Energies Prediction on Bimetallic Transition Metal Surfaces Using an Interpretable Machine Learning-Accelerated Density Functional Theory Approach

Atomic Adsorption Energies Prediction on Bimetallic Transition Metal Surfaces Using an Interpretable Machine Learning-Accelerated Density Functional Theory Approach

In this study, we identified features with the largest contributions and property trends in predicting the adsorption energies of carbon, hydrogen, and oxygen adsorbates on transition metal (TM) surfaces by performing Density Functional Theory (DFT)-based calculations and Machine Learning (ML) regression models. From 26 monometallic and 400 bimetallic fcc(111) TM surfaces obtained from Catalysis-hub.org, three datasets consisting of fourteen elemental, electronic, and structural properties were generated using DFT calculations, site calculations, and online databases. The number of features was reduced using feature selection and then finely-tuned random forest regression (RFR), gaussian process regression (GPR), and artificial neural network (ANN) algorithms were implemented for adsorption energy prediction. Finally, model-agnostic interpretation methods such as permutation feature importance (PFI) and shapely additive explanations (SHAP) provided rankings of feature contributions and directional trends. For all datasets, RFR and GPR demonstrated the highest prediction accuracies. In addition, interpretation methods demonstrated that the largest contributing features and directional trends in the regression models were consistent with structure-property-performance relationships of TMs like the d-band model, the Friedel model, and higher-fold adsorption sites. Overall, this interpretable ML–DFT approach can be applied to TMs and their derivatives for atomic adsorption energy prediction and model explainability.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
ChemistryOpen
ChemistryOpen CHEMISTRY, MULTIDISCIPLINARY-
CiteScore
4.80
自引率
4.30%
发文量
143
审稿时长
1 months
期刊介绍: ChemistryOpen is a multidisciplinary, gold-road open-access, international forum for the publication of outstanding Reviews, Full Papers, and Communications from all areas of chemistry and related fields. It is co-owned by 16 continental European Chemical Societies, who have banded together in the alliance called ChemPubSoc Europe for the purpose of publishing high-quality journals in the field of chemistry and its border disciplines. As some of the governments of the countries represented in ChemPubSoc Europe have strongly recommended that the research conducted with their funding is freely accessible for all readers (Open Access), ChemPubSoc Europe was concerned that no journal for which the ethical standards were monitored by a chemical society was available for such papers. ChemistryOpen fills this gap.
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信