Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost

Band selection (BS) identifies key bands from hyperspectral imagery (HSI) for specific downstream tasks, playing a pivotal role in practical applications. eXtreme Gradient Boosting (XGBoost), an interpretable tree-based ensemble learning classifier, explicitly implements the complex nonlinear hypers...

Full description

Saved in:
Bibliographic Details
Main Authors: Xiaodi Shang, Chuanyu Cui, Xudong Sun, Xiaopeng Wang, Jiahua Zhang
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11008687/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849435309474840576
author Xiaodi Shang
Chuanyu Cui
Xudong Sun
Xiaopeng Wang
Jiahua Zhang
author_facet Xiaodi Shang
Chuanyu Cui
Xudong Sun
Xiaopeng Wang
Jiahua Zhang
author_sort Xiaodi Shang
collection DOAJ
description Band selection (BS) identifies key bands from hyperspectral imagery (HSI) for specific downstream tasks, playing a pivotal role in practical applications. eXtreme Gradient Boosting (XGBoost), an interpretable tree-based ensemble learning classifier, explicitly implements the complex nonlinear hyperspectral classification. The interpretable information extracted from the tree structure offers a novel basis for supervised BS. To this end, this article proposes a supervised BS method, named classification task-driven hyperspectral BS via interpretability from XGBoost (XGBS). It leverages prior knowledge to train a classification task-driven XGBoost and interprets the tree structure to extract multivariate interpretable information, encompassing band split gain and two types of band dependencies. Subsequently, a heuristic search framework is employed to evaluate band performance, facilitating the selection of bands with strong classification capability, high interdependency, and low redundancy. Experiments conducted on six real HSI datasets demonstrate the effectiveness and stability of the proposed XGBS.
format Article
id doaj-art-8f3f1a9cd4264366b0a11b97a9cf9b1d
institution Kabale University
issn 1939-1404
2151-1535
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
spelling doaj-art-8f3f1a9cd4264366b0a11b97a9cf9b1d2025-08-20T03:26:20ZengIEEEIEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing1939-14042151-15352025-01-0118137331375410.1109/JSTARS.2025.357227811008687Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoostXiaodi Shang0https://orcid.org/0000-0002-0133-8447Chuanyu Cui1https://orcid.org/0009-0000-5479-6938Xudong Sun2https://orcid.org/0000-0002-5870-6343Xiaopeng Wang3https://orcid.org/0009-0005-8965-7758Jiahua Zhang4https://orcid.org/0000-0002-2894-9627College of Computer Science and Technology, Qingdao University, Qingdao, ChinaCollege of Computer Science and Technology, Qingdao University, Qingdao, ChinaCollege of Computer Science and Technology, Qingdao University, Qingdao, ChinaCollege of Computer Science and Technology, Qingdao University, Qingdao, ChinaAerospace Information Research Institute, Chinese Academy of Sciences, Beijing, ChinaBand selection (BS) identifies key bands from hyperspectral imagery (HSI) for specific downstream tasks, playing a pivotal role in practical applications. eXtreme Gradient Boosting (XGBoost), an interpretable tree-based ensemble learning classifier, explicitly implements the complex nonlinear hyperspectral classification. The interpretable information extracted from the tree structure offers a novel basis for supervised BS. To this end, this article proposes a supervised BS method, named classification task-driven hyperspectral BS via interpretability from XGBoost (XGBS). It leverages prior knowledge to train a classification task-driven XGBoost and interprets the tree structure to extract multivariate interpretable information, encompassing band split gain and two types of band dependencies. Subsequently, a heuristic search framework is employed to evaluate band performance, facilitating the selection of bands with strong classification capability, high interdependency, and low redundancy. Experiments conducted on six real HSI datasets demonstrate the effectiveness and stability of the proposed XGBS.https://ieeexplore.ieee.org/document/11008687/Band dependenceclassification task-driven band selection (BS)eXtreme Gradient Boosting (XGBoost)hyperspectral imagery (HSI)multivariate interpretable information (MII)
spellingShingle Xiaodi Shang
Chuanyu Cui
Xudong Sun
Xiaopeng Wang
Jiahua Zhang
Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing
Band dependence
classification task-driven band selection (BS)
eXtreme Gradient Boosting (XGBoost)
hyperspectral imagery (HSI)
multivariate interpretable information (MII)
title Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
title_full Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
title_fullStr Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
title_full_unstemmed Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
title_short Classification Task-Driven Hyperspectral Band Selection via Interpretability From XGBoost
title_sort classification task driven hyperspectral band selection via interpretability from xgboost
topic Band dependence
classification task-driven band selection (BS)
eXtreme Gradient Boosting (XGBoost)
hyperspectral imagery (HSI)
multivariate interpretable information (MII)
url https://ieeexplore.ieee.org/document/11008687/
work_keys_str_mv AT xiaodishang classificationtaskdrivenhyperspectralbandselectionviainterpretabilityfromxgboost
AT chuanyucui classificationtaskdrivenhyperspectralbandselectionviainterpretabilityfromxgboost
AT xudongsun classificationtaskdrivenhyperspectralbandselectionviainterpretabilityfromxgboost
AT xiaopengwang classificationtaskdrivenhyperspectralbandselectionviainterpretabilityfromxgboost
AT jiahuazhang classificationtaskdrivenhyperspectralbandselectionviainterpretabilityfromxgboost