A Geospatial and Machine Learning Approach
Suresh Mondal,Mst Tania Parveen,Asraful Alam,Rukhsana,Nazrul Islam,Beata Calka,Bashar Bashir,Mohamed Zhran
DOI: https://doi.org/10.3390/ijgi13110388
IF: 3.4
2024-10-31
ISPRS International Journal of Geo-Information
Abstract:The rapid urbanization occurring globally has significantly intensified the challenges of waste management in densely populated metropolitan areas. A growing amount of waste has become a major concern for municipal authorities and local governments due to the limited availability of suitable land. Geospatial techniques, such as Geographic Information Systems (GISs) and remote sensing, combined with machine learning, play a crucial role in identifying suitable sites for urban waste management. These techniques assist planners in making well-informed decisions that strike a balance between environmental preservation and urban expansion by examining spatial data on land use, population density, and environmental concerns. Geospatial tools provide a data-driven basis for policy and urban planning, ensuring effective land use, reducing ecological hazards, and promoting sustainable urban growth for municipalities such as English Bazar and Old Malda. It can also pose serious threats to the environment, public health, and communities. Focusing on the English Bazar and Old Malda Municipalities in India, this paper examines the use of geospatial technologies to identify suitable sites for waste disposal. The research aims to address the complex processes of waste generation, collection, and disposal in urban environments. Using GIS and a Multi-Criteria Decision Analysis (MCDA) approach, the study employs the Analytic Hierarchy Process (AHP) alongside the Random Forest (RF) model and a machine learning (ML) technique to identify potential waste disposal sites within the English Bazar and Old Malda Municipalities in the Malda district. Eight key criteria were considered in the site selection process: land elevation; distances from surface water, roads, railways, and urban areas; groundwater depth; land use and land cover; and distance from sensitive and restricted areas. AHP analysis showed that 8%, 26%, and 27% of the sites were categorized as very highly suitable, moderately suitable, and unsuitable, respectively. Meanwhile, 38%, 17%, and 13% of the areas were classified as unsuitable, moderately suitable, and very highly suitable according to the RF model. The overall accuracy and Kappa coefficient indicated that the AHP method (overall capacity of 83.83% and Kappa coefficient of 0.7894) was slightly better than the RF model (overall capacity of 80.61% and Kappa coefficient of 0.7474) for site suitability analysis. This research underscores the broad relevance of geospatial technology in creating resilient and environmentally sustainable cities while offering valuable guidance on effectively allocating waste disposal sites. The findings provide crucial insights for urban planners and decision-makers, facilitating the identification of optimal locations for sustainable waste management in urban settings.
geography, physical,remote sensing,computer science, information systems