Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application
This study evaluates the impact of four feature selection (FS) algorithms in an object-based image analysis framework for very-high-resolution land use-land cover classification. The selected FS algorithms, correlation-based feature selection, mean decrease in accuracy, random forest (RF) based recu...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2018-03-01
|
| Series: | GIScience & Remote Sensing |
| Subjects: | |
| Online Access: | http://dx.doi.org/10.1080/15481603.2017.1408892 |
| _version_ | 1827811799812014080 |
|---|---|
| author | Stefanos Georganos Tais Grippa Sabine Vanhuysse Moritz Lennert Michal Shimoni Stamatis Kalogirou Eleonore Wolff |
| author_facet | Stefanos Georganos Tais Grippa Sabine Vanhuysse Moritz Lennert Michal Shimoni Stamatis Kalogirou Eleonore Wolff |
| author_sort | Stefanos Georganos |
| collection | DOAJ |
| description | This study evaluates the impact of four feature selection (FS) algorithms in an object-based image analysis framework for very-high-resolution land use-land cover classification. The selected FS algorithms, correlation-based feature selection, mean decrease in accuracy, random forest (RF) based recursive feature elimination, and variable selection using random forest, were tested on the extreme gradient boosting, support vector machine, K-nearest neighbor, RF, and recursive partitioningclassifiers, respectively. The results demonstrate that the selection of an appropriate FS method can be crucial to the performance of a machine learning classifier in terms of accuracy but also parsimony. In this scope, we propose a new metric to perform model selection named classification optimization score (COS) that rewards model simplicity and indirectly penalizes for increased computational time and processing requirements using the number of features for a given classification model as a surrogate. Our findings suggest that applying rigorous FS along with utilizing the COS metric may significantly reduce the processing time and the storage space while at the same time producing higher classification accuracy than using the initial dataset. |
| first_indexed | 2024-03-11T23:09:45Z |
| format | Article |
| id | doaj.art-bfaeb9b3c9464fd0b26ad3e75ed49db0 |
| institution | Directory Open Access Journal |
| issn | 1548-1603 1943-7226 |
| language | English |
| last_indexed | 2024-03-11T23:09:45Z |
| publishDate | 2018-03-01 |
| publisher | Taylor & Francis Group |
| record_format | Article |
| series | GIScience & Remote Sensing |
| spelling | doaj.art-bfaeb9b3c9464fd0b26ad3e75ed49db02023-09-21T12:34:14ZengTaylor & Francis GroupGIScience & Remote Sensing1548-16031943-72262018-03-0155222124210.1080/15481603.2017.14088921408892Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban applicationStefanos Georganos0Tais Grippa1Sabine Vanhuysse2Moritz Lennert3Michal Shimoni4Stamatis Kalogirou5Eleonore Wolff6Universite Libre de BruxellesUniversite Libre de BruxellesUniversite Libre de BruxellesUniversite Libre de BruxellesRoyal Military AcademyHarokopio University of AthensUniversite Libre de BruxellesThis study evaluates the impact of four feature selection (FS) algorithms in an object-based image analysis framework for very-high-resolution land use-land cover classification. The selected FS algorithms, correlation-based feature selection, mean decrease in accuracy, random forest (RF) based recursive feature elimination, and variable selection using random forest, were tested on the extreme gradient boosting, support vector machine, K-nearest neighbor, RF, and recursive partitioningclassifiers, respectively. The results demonstrate that the selection of an appropriate FS method can be crucial to the performance of a machine learning classifier in terms of accuracy but also parsimony. In this scope, we propose a new metric to perform model selection named classification optimization score (COS) that rewards model simplicity and indirectly penalizes for increased computational time and processing requirements using the number of features for a given classification model as a surrogate. Our findings suggest that applying rigorous FS along with utilizing the COS metric may significantly reduce the processing time and the storage space while at the same time producing higher classification accuracy than using the initial dataset.http://dx.doi.org/10.1080/15481603.2017.1408892obialand cover classificationextreme gradient boostingfeature selectionmachine learning |
| spellingShingle | Stefanos Georganos Tais Grippa Sabine Vanhuysse Moritz Lennert Michal Shimoni Stamatis Kalogirou Eleonore Wolff Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application GIScience & Remote Sensing obia land cover classification extreme gradient boosting feature selection machine learning |
| title | Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application |
| title_full | Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application |
| title_fullStr | Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application |
| title_full_unstemmed | Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application |
| title_short | Less is more: optimizing classification performance through feature selection in a very-high-resolution remote sensing object-based urban application |
| title_sort | less is more optimizing classification performance through feature selection in a very high resolution remote sensing object based urban application |
| topic | obia land cover classification extreme gradient boosting feature selection machine learning |
| url | http://dx.doi.org/10.1080/15481603.2017.1408892 |
| work_keys_str_mv | AT stefanosgeorganos lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT taisgrippa lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT sabinevanhuysse lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT moritzlennert lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT michalshimoni lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT stamatiskalogirou lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication AT eleonorewolff lessismoreoptimizingclassificationperformancethroughfeatureselectioninaveryhighresolutionremotesensingobjectbasedurbanapplication |