EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

Comparative assessment of flood risk using AHP and machine learning models in Kurigram district, Bangladesh

Pranta Bhowmik, Aminul Haque Laskor, Ananta Bhowmik, Mohaimenul Islam

Journal of Geospatial Science and Analytics · 30.07.2026 · 2 (2) · 159-186

DOI: 10.58524/jgsa.v2i2.139 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Geospatial Science and Analytics, online erschienen am 30.07.2026. Verfasst von Pranta Bhowmik und 3 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Flood risk assessment requires an integrated evaluation of physical flood hazards and socio-economic vulnerability, particularly in highly flood-prone regions such as northern Bangladesh. This study presents a GIS-based flood risk assessment framework integrating the Analytic Hierarchy Process (AHP) and machine learning models for spatial flood risk mapping in Kurigram District, Bangladesh. The novelty of the study lies in combining environmental flood-hazard factors and socio-economic vulnerability indicators within a hybrid GIS–AHP–ML framework for comprehensive flood risk assessment. A Flood Hazard Index (FHI) was developed using ten physical factors, while socio-economic vulnerability was assessed through a Flood Vulnerability Index (FVI). The AHP-derived weights produced acceptable consistency ratios of 0.08 for FHI and 0.014 for FVI, confirming the reliability of factor prioritization. The Flood Risk Index (FRI) was generated by integrating FHI and FVI within a GIS environment and classified into five risk categories. Results indicate that flood risk in Kurigram District is influenced more strongly by socio-economic vulnerability and limited adaptive capacity than by physical hazard intensity alone. Among the applied machine learning models, Random Forest achieved the highest predictive performance and identified 173.28 km² of Nageshwari Upazila as high-risk. Model validation using Random Forest, Support Vector Machine, and Decision Tree produced AUC values of 0.93, 0.91, and 0.86, respectively, compared to 0.82 for the conventional AHP approach. The findings demonstrate that the proposed hybrid GIS–AHP–ML framework significantly improves flood risk prediction accuracy and provides a transferable, policy-relevant tool for disaster risk reduction, land-use planning, and climate resilience building in flood-prone regions.

Zitieren

APA

Bhowmik, P., Laskor, A. H., Bhowmik, A., & Islam, M. (2026). Comparative assessment of flood risk using AHP and machine learning models in Kurigram district, Bangladesh. Journal of Geospatial Science and Analytics, 2(2), 159-186. https://doi.org/10.58524/jgsa.v2i2.139

Vancouver

Bhowmik P, Laskor AH, Bhowmik A, Islam M. Comparative assessment of flood risk using AHP and machine learning models in Kurigram district, Bangladesh. Journal of Geospatial Science and Analytics. 2026;2(2):159-186. doi:10.58524/jgsa.v2i2.139

BibTeX
@article{Bhowmik2026,
  author = {Bhowmik, Pranta and Laskor, Aminul Haque and Bhowmik, Ananta and Islam, Mohaimenul},
  title = {Comparative assessment of flood risk using AHP and machine learning models in Kurigram district, Bangladesh},
  journal = {Journal of Geospatial Science and Analytics},
  year = {2026},
  volume = {2},
  number = {2},
  pages = {159--186},
  doi = {10.58524/jgsa.v2i2.139}
}
Kostenlos starten →