High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction

Thermoelectric materials, which convert thermal energy directly into electricity, hold promise for sustainable energy applications. However, accurately predicting their efficiency, quantified by the figure of merit (zT), remains challenging, especially for doped materials. Here we present a machine...

Full description

Saved in:
Bibliographic Details
Main Authors: Yuelin Wang, Chengquan Zhong, Jingzi Zhang, Honghao Yao, Junjie Chen, Xi Lin
Format: Article
Language:English
Published: Elsevier 2025-01-01
Series:Materials & Design
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S0264127524009274
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1841553897583280128
author Yuelin Wang
Chengquan Zhong
Jingzi Zhang
Honghao Yao
Junjie Chen
Xi Lin
author_facet Yuelin Wang
Chengquan Zhong
Jingzi Zhang
Honghao Yao
Junjie Chen
Xi Lin
author_sort Yuelin Wang
collection DOAJ
description Thermoelectric materials, which convert thermal energy directly into electricity, hold promise for sustainable energy applications. However, accurately predicting their efficiency, quantified by the figure of merit (zT), remains challenging, especially for doped materials. Here we present a machine learning (ML) approach, the stacking model, that significantly improves zT prediction accuracy for doped thermoelectric. By combining five regression models through stacking ensemble learning and introducing 100 coordination number features alongside conventional features, our model achieves a coefficient of determination (R2) value of 0.970. This high performance demonstrates unprecedented sensitivity to zT variations due to doping. We validate our model using an expanded dataset of over 230 new materials from recent literature. The model identifies 43 potential high-zT materials, including Pb0.97K0.03Te0.65S0.25Se0.1 with a predicted zT of 1.9. Density functional theory calculations confirm the superior electrical properties of this compound. Our approach offers an efficient strategy for large-scale screening of high-performance thermoelectric materials, potentially accelerating their discovery and development for energy applications.
format Article
id doaj-art-7ae81134613d40ff871fdfb40afb7390
institution Kabale University
issn 0264-1275
language English
publishDate 2025-01-01
publisher Elsevier
record_format Article
series Materials & Design
spelling doaj-art-7ae81134613d40ff871fdfb40afb73902025-01-09T06:12:23ZengElsevierMaterials & Design0264-12752025-01-01249113552High-Performance stacking ensemble learning for thermoelectric figure-of-merit predictionYuelin Wang0Chengquan Zhong1Jingzi Zhang2Honghao Yao3Junjie Chen4Xi Lin5School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Blockchain Development and Research Institute, Harbin Institute of Technology, Shenzhen 518055, Guangdong, ChinaSchool of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Blockchain Development and Research Institute, Harbin Institute of Technology, Shenzhen 518055, Guangdong, ChinaSchool of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Blockchain Development and Research Institute, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Corresponding authors at: School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China.School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Blockchain Development and Research Institute, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Corresponding authors at: School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China.School of Computer Science and Technology, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Corresponding author.School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, Heilongjiang, China; Blockchain Development and Research Institute, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China; Corresponding authors at: School of Materials Science and Engineering, Harbin Institute of Technology, Shenzhen 518055, Guangdong, China.Thermoelectric materials, which convert thermal energy directly into electricity, hold promise for sustainable energy applications. However, accurately predicting their efficiency, quantified by the figure of merit (zT), remains challenging, especially for doped materials. Here we present a machine learning (ML) approach, the stacking model, that significantly improves zT prediction accuracy for doped thermoelectric. By combining five regression models through stacking ensemble learning and introducing 100 coordination number features alongside conventional features, our model achieves a coefficient of determination (R2) value of 0.970. This high performance demonstrates unprecedented sensitivity to zT variations due to doping. We validate our model using an expanded dataset of over 230 new materials from recent literature. The model identifies 43 potential high-zT materials, including Pb0.97K0.03Te0.65S0.25Se0.1 with a predicted zT of 1.9. Density functional theory calculations confirm the superior electrical properties of this compound. Our approach offers an efficient strategy for large-scale screening of high-performance thermoelectric materials, potentially accelerating their discovery and development for energy applications.http://www.sciencedirect.com/science/article/pii/S0264127524009274Thermoelectric materialszTMachine LearningStacking ensemble model
spellingShingle Yuelin Wang
Chengquan Zhong
Jingzi Zhang
Honghao Yao
Junjie Chen
Xi Lin
High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
Materials & Design
Thermoelectric materials
zT
Machine Learning
Stacking ensemble model
title High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
title_full High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
title_fullStr High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
title_full_unstemmed High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
title_short High-Performance stacking ensemble learning for thermoelectric figure-of-merit prediction
title_sort high performance stacking ensemble learning for thermoelectric figure of merit prediction
topic Thermoelectric materials
zT
Machine Learning
Stacking ensemble model
url http://www.sciencedirect.com/science/article/pii/S0264127524009274
work_keys_str_mv AT yuelinwang highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction
AT chengquanzhong highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction
AT jingzizhang highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction
AT honghaoyao highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction
AT junjiechen highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction
AT xilin highperformancestackingensemblelearningforthermoelectricfigureofmeritprediction