Disruption prediction and model analysis using LightGBM on J-TEXT and HL-2A

doi:10.1088/1361-6587/abfa74

	Disruption prediction and model analysis using LightGBM on J-TEXT and HL-2A
	Zhong,Y 2; Zheng,W 2; Chen,Z Y 1,2; Xia,F 3; Yu,L M 3; Wu,Q Q 2; Ai,X K 2; Shen,C S 2; Yang,Z Y 3; Yan,W 2
刊名	Plasma Physics and Controlled Fusion
	2021-05-21
卷号	63
关键词	disruption prediction machine learning LightGBM feature importance
ISSN号	0741-3335
DOI	10.1088/1361-6587/abfa74
通讯作者	Zheng,W() ; Chen,Z Y()
英文摘要	AbstractUsing machine learning (ML) techniques to develop disruption predictors is an effective way to avoid or mitigate the disruption in a large-scale tokamak. The recent ML-based disruption predictors have made great progress regarding accuracy, but most of them have not achieved acceptable cross-machine performance. Before we develop a cross-machine predictor, it is very important to investigate the method of developing a cross-tokamak ML-based disruption prediction model. To ascertain the elements which impact the model’s performance and achieve a deep understanding of the predictor, multiple models are trained using data from two different tokamaks, J-TEXT and HL-2A, based on an implementation of the gradient-boosted decision trees algorithm called LightGBM, which can provide detailed information about the model and input features. The predictor models are not only built and tested for performance, but also analyzed from a feature importance perspective as well as for model performance variation. The relative feature importance ranking of two tokamaks is caused by differences in disruption types between different tokamaks. The result of two models with seven inputs showed that common diagnostics is very important in building a cross-machine predictor. This provided a strategy for selecting diagnostics and shots data for developing cross-machine predictors.
资助项目	National Key Research and Development Program of China[2017YFE0301202] ; National Key Research and Development Program of China[2017YFE0300500] ; National Key Research and Development Program of China[2017YFE0300501] ; National Magnetic Confinement Fusion Science Program[2015GB111002] ; National Magnetic Confinement Fusion Science Program[2015GB104000] ; National Natural Science Foundation of China[11775089] ; National Natural Science Foundation of China[51821005] ; National Natural Science Foundation of China[71762031] ; National Natural Science Foundation of China[11875022] ; National Natural Science Foundation of China[11905077] ; National Natural Science Foundation of China[11575068] ; National MCF Energy R&D Program of China[2019YFE03010004] ; China Postdoctoral Science Foundation[2019M652615]
WOS关键词	TOOL
WOS研究方向	Physics
语种	英语
出版者	IOP Publishing
WOS记录号	IOP:0741-3335-63-7-ABFA74
资助机构	National Key Research and Development Program of China ; National Magnetic Confinement Fusion Science Program ; National Natural Science Foundation of China ; National MCF Energy R&D Program of China ; China Postdoctoral Science Foundation
内容类型	期刊论文
源URL	[http://ir.hfcas.ac.cn:8080/handle/334002/122528]
专题	中国科学院合肥物质科学研究院
通讯作者	Zheng,W; Chen,Z Y
作者单位	1.Chengdu University, 610106 Chengdu, People’s Republic of China 2.International Joint Research Laboratory of Magnetic Confinement Fusion and Plasma Physics, State Key Laboratory of Advanced Electromagnetic Engineering and Technology, School of Electrical and Electronic Engineering, Huazhong University of Science and Technology, Wuhan 430074, People’s Republic of China 3.Southwestern Institute of Physics, Chengdu 610041, People’s Republic of China 4.Forschungszentrum Jülich GmbH, Institut für Energie- und Klimaforschung—Plasmaphysik, 52425 Jülich, Germany 5.Institute of Plasma Physics, Chinese Academy of Sciences, PO Box 1126, Hefei 230031, People’s Republic of China
推荐引用方式 GB/T 7714	Zhong,Y,Zheng,W,Chen,Z Y,et al. Disruption prediction and model analysis using LightGBM on J-TEXT and HL-2A[J]. Plasma Physics and Controlled Fusion,2021,63.
APA	Zhong,Y.,Zheng,W.,Chen,Z Y.,Xia,F.,Yu,L M.,...&Li,F.(2021).Disruption prediction and model analysis using LightGBM on J-TEXT and HL-2A.Plasma Physics and Controlled Fusion,63.
MLA	Zhong,Y,et al."Disruption prediction and model analysis using LightGBM on J-TEXT and HL-2A".Plasma Physics and Controlled Fusion 63(2021).