Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones

Every year, millions of individuals are affected by flooding in Bangladesh resulting in loss of habitat, properties, and lives. Because the country experiences monsoon rains, flat lands, cyclones as well as glacier melt, it is highly prone to flooding. This work aimed to lessen the impact of floodin...

Full description

Saved in:
Bibliographic Details
Main Authors: Md Raqibul Hasan, Md. Jamal Hossain, Md. Waliullah, Abdul Hannan, Md. Mijanoor Rahman
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/11052287/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850087760694607872
author Md Raqibul Hasan
Md. Jamal Hossain
Md. Waliullah
Abdul Hannan
Md. Mijanoor Rahman
author_facet Md Raqibul Hasan
Md. Jamal Hossain
Md. Waliullah
Abdul Hannan
Md. Mijanoor Rahman
author_sort Md Raqibul Hasan
collection DOAJ
description Every year, millions of individuals are affected by flooding in Bangladesh resulting in loss of habitat, properties, and lives. Because the country experiences monsoon rains, flat lands, cyclones as well as glacier melt, it is highly prone to flooding. This work aimed to lessen the impact of flooding disasters and develop effective strategies for early warning systems. The intertwined weather patterns known as the atmospheric river (AR) of Bangladesh make use of topological data analysis (TDA) in connection with wavelet decomposition and unsupervised machine learning (k-means clustering) methods to pave the way for enhanced flood detection. We have pinpointed the most desired cluster group, which stands at k =4 in our argument. In terms of the silhouette coefficient, this group had the highest values in Rangpur at 0.9016, Sylhet at 0.9028, Barisal at 0.9094, and Khulna at 0.9014. Our work was highly efficient and ranked at approximately 96% using SVMs. Later, we identified flooded and non-flooded regions with the same technique. The potential flood-prone periods were identified for all four regions based on optimal clustering results, as indicated by silhouette coefficient values. At any time when there is a possibility of flood occurrence, it can alert residents in the affected regions to move out to safer grounds; this could also stimulate appropriate actions from the risk managers.
format Article
id doaj-art-64b8aa5706e14aa1a3a40b0b96c14700
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-64b8aa5706e14aa1a3a40b0b96c147002025-08-20T02:43:10ZengIEEEIEEE Access2169-35362025-01-011311171011172110.1109/ACCESS.2025.358358111052287Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding ZonesMd Raqibul Hasan0Md. Jamal Hossain1https://orcid.org/0000-0003-3093-0033Md. Waliullah2Abdul Hannan3Md. Mijanoor Rahman4https://orcid.org/0009-0000-8344-5432Department of Applied Mathematics, Noakhali Science and Technology University, Noakhali, BangladeshDepartment of Applied Mathematics, Noakhali Science and Technology University, Noakhali, BangladeshDepartment of Applied Mathematics, Noakhali Science and Technology University, Noakhali, BangladeshDepartment of Applied Mathematics, Noakhali Science and Technology University, Noakhali, Bangladesh2Department of Mathematics, Mawlana Bhashani Science and Technology University, Santosh, Tangail, BangladeshEvery year, millions of individuals are affected by flooding in Bangladesh resulting in loss of habitat, properties, and lives. Because the country experiences monsoon rains, flat lands, cyclones as well as glacier melt, it is highly prone to flooding. This work aimed to lessen the impact of flooding disasters and develop effective strategies for early warning systems. The intertwined weather patterns known as the atmospheric river (AR) of Bangladesh make use of topological data analysis (TDA) in connection with wavelet decomposition and unsupervised machine learning (k-means clustering) methods to pave the way for enhanced flood detection. We have pinpointed the most desired cluster group, which stands at k =4 in our argument. In terms of the silhouette coefficient, this group had the highest values in Rangpur at 0.9016, Sylhet at 0.9028, Barisal at 0.9094, and Khulna at 0.9014. Our work was highly efficient and ranked at approximately 96% using SVMs. Later, we identified flooded and non-flooded regions with the same technique. The potential flood-prone periods were identified for all four regions based on optimal clustering results, as indicated by silhouette coefficient values. At any time when there is a possibility of flood occurrence, it can alert residents in the affected regions to move out to safer grounds; this could also stimulate appropriate actions from the risk managers.https://ieeexplore.ieee.org/document/11052287/Floodatmospheric rivertopological data analysiswaveletunsupervised machine learningk-means clustering
spellingShingle Md Raqibul Hasan
Md. Jamal Hossain
Md. Waliullah
Abdul Hannan
Md. Mijanoor Rahman
Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
IEEE Access
Flood
atmospheric river
topological data analysis
wavelet
unsupervised machine learning
k-means clustering
title Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
title_full Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
title_fullStr Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
title_full_unstemmed Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
title_short Topological Data Analysis and Wavelet- Unsupervised Machine Learning Approaches to Identifying the Flooding and Non-Flooding Zones
title_sort topological data analysis and wavelet unsupervised machine learning approaches to identifying the flooding and non flooding zones
topic Flood
atmospheric river
topological data analysis
wavelet
unsupervised machine learning
k-means clustering
url https://ieeexplore.ieee.org/document/11052287/
work_keys_str_mv AT mdraqibulhasan topologicaldataanalysisandwaveletunsupervisedmachinelearningapproachestoidentifyingthefloodingandnonfloodingzones
AT mdjamalhossain topologicaldataanalysisandwaveletunsupervisedmachinelearningapproachestoidentifyingthefloodingandnonfloodingzones
AT mdwaliullah topologicaldataanalysisandwaveletunsupervisedmachinelearningapproachestoidentifyingthefloodingandnonfloodingzones
AT abdulhannan topologicaldataanalysisandwaveletunsupervisedmachinelearningapproachestoidentifyingthefloodingandnonfloodingzones
AT mdmijanoorrahman topologicaldataanalysisandwaveletunsupervisedmachinelearningapproachestoidentifyingthefloodingandnonfloodingzones