A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models
Unaddressed imbalance of multitemporal land cover (LC) data reduces deep learning (DL) model usefulness to forecast changes. To manage geospatial data imbalance, there is a lack of specialized cost-sensitive learning strategies available. Sample weights are typically derived from training instance f...
Main Authors: | , |
---|---|
Format: | Article |
Language: | English |
Published: |
Taylor & Francis Group
2023-12-01
|
Series: | Geocarto International |
Subjects: | |
Online Access: | http://dx.doi.org/10.1080/10106049.2023.2240283 |
_version_ | 1797681566675107840 |
---|---|
author | Alysha van Duynhoven Suzana Dragićević |
author_facet | Alysha van Duynhoven Suzana Dragićević |
author_sort | Alysha van Duynhoven |
collection | DOAJ |
description | Unaddressed imbalance of multitemporal land cover (LC) data reduces deep learning (DL) model usefulness to forecast changes. To manage geospatial data imbalance, there is a lack of specialized cost-sensitive learning strategies available. Sample weights are typically derived from training instance frequencies which disregard spatial pattern complexities. Therefore, this study proposes a geospatial sample weighting approach underpinned by class-level landscape metrics (LSMs) to assign importance to categories based on relative indicators of spatial form. A case study demonstrates the application and effects of the LSM-based sample weighting approach for projecting LC changes of a region in British Columbia, Canada. Four spatiotemporal DL models are provided weighted training samples including multitemporal LC data and explanatory factors. Sample weights calculated from indicators of patch density, shape irregularity and shape heterogeneity improved figure of merit and related measures over baseline configurations. This study contributes to LC change data imbalance management strategies for DL models. |
first_indexed | 2024-03-11T23:46:45Z |
format | Article |
id | doaj.art-20392b2f476f4f29855c2422dde3286b |
institution | Directory Open Access Journal |
issn | 1010-6049 1752-0762 |
language | English |
last_indexed | 2024-03-11T23:46:45Z |
publishDate | 2023-12-01 |
publisher | Taylor & Francis Group |
record_format | Article |
series | Geocarto International |
spelling | doaj.art-20392b2f476f4f29855c2422dde3286b2023-09-19T09:13:18ZengTaylor & Francis GroupGeocarto International1010-60491752-07622023-12-0138110.1080/10106049.2023.22402832240283A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning modelsAlysha van Duynhoven0Suzana Dragićević1Spatial Analysis and Modeling Laboratory, Department of Geography, Simon Fraser UniversitySpatial Analysis and Modeling Laboratory, Department of Geography, Simon Fraser UniversityUnaddressed imbalance of multitemporal land cover (LC) data reduces deep learning (DL) model usefulness to forecast changes. To manage geospatial data imbalance, there is a lack of specialized cost-sensitive learning strategies available. Sample weights are typically derived from training instance frequencies which disregard spatial pattern complexities. Therefore, this study proposes a geospatial sample weighting approach underpinned by class-level landscape metrics (LSMs) to assign importance to categories based on relative indicators of spatial form. A case study demonstrates the application and effects of the LSM-based sample weighting approach for projecting LC changes of a region in British Columbia, Canada. Four spatiotemporal DL models are provided weighted training samples including multitemporal LC data and explanatory factors. Sample weights calculated from indicators of patch density, shape irregularity and shape heterogeneity improved figure of merit and related measures over baseline configurations. This study contributes to LC change data imbalance management strategies for DL models.http://dx.doi.org/10.1080/10106049.2023.2240283land cover changelandscape metricsgeographic data imbalancesample weightsspatiotemporal deep learning |
spellingShingle | Alysha van Duynhoven Suzana Dragićević A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models Geocarto International land cover change landscape metrics geographic data imbalance sample weights spatiotemporal deep learning |
title | A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models |
title_full | A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models |
title_fullStr | A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models |
title_full_unstemmed | A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models |
title_short | A landscape metrics-based sample weighting approach for forecasting land cover change with deep learning models |
title_sort | landscape metrics based sample weighting approach for forecasting land cover change with deep learning models |
topic | land cover change landscape metrics geographic data imbalance sample weights spatiotemporal deep learning |
url | http://dx.doi.org/10.1080/10106049.2023.2240283 |
work_keys_str_mv | AT alyshavanduynhoven alandscapemetricsbasedsampleweightingapproachforforecastinglandcoverchangewithdeeplearningmodels AT suzanadragicevic alandscapemetricsbasedsampleweightingapproachforforecastinglandcoverchangewithdeeplearningmodels AT alyshavanduynhoven landscapemetricsbasedsampleweightingapproachforforecastinglandcoverchangewithdeeplearningmodels AT suzanadragicevic landscapemetricsbasedsampleweightingapproachforforecastinglandcoverchangewithdeeplearningmodels |