Graph Neural Network Based Modelling of Pest Spread Across Spatially Connected Agricultural Fields
Journal of Smart and Sustainable Farming · 26.04.2026 · 062
DOI: 10.64026/jssf/2026007 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Smart and Sustainable Farming, online erschienen am 26.04.2026. Verfasst von Shweta Sharma. Die Arbeit nennt 16 Literaturangaben.
Abstract (Originalsprache)
Pest infestation has been a significant threat to agricultural productivity because the spread of these pests could have been possible from neighbouring fields by spatial connectivity, environmental similarity, and crop-to-crop interactions. The spatial dependenci es were previously under-represented and early infestation detection was limited in the traditional pest prediction methods used in agriculture. A pest transmission model, called a Spatial Graph Neural Network for Pest Spread Prediction (SGNN-PSP) had been suggested to predict the spatial spread of a pest in an agricultural field network. Agricultural fields were modeled as the graph nodes, and the spatial adjacency, geographical closeness and connectivity relationship between fields were modeled as the graph edges. The framework proposed had combined graph convolution with temporal feature learning to model both spatial propagation patterns and temporal changes in pest activity. Field level environmental, crop and historical infestation characteristics were included to model pest spread risk within contiguous fields. Experimental results have shown that SGNN-PSP outperforms conventional machine-learning and deep-learning methods in terms of the pest-risk prediction. The model had been able to better identify high-risk fields, better spatial generalize, and earlier recognize potential propagations of infestation. The results showed that graph-based spatial modelling could aid proactive pest surveillance and agricultural intervention at a more specific level, while limiting the number of treatments applied within the field.
Zitieren
APA
Sharma, S. (2026). Graph Neural Network Based Modelling of Pest Spread Across Spatially Connected Agricultural Fields. Journal of Smart and Sustainable Farming, 062. https://doi.org/10.64026/jssf/2026007
Vancouver
Sharma S. Graph Neural Network Based Modelling of Pest Spread Across Spatially Connected Agricultural Fields. Journal of Smart and Sustainable Farming. 2026:062. doi:10.64026/jssf/2026007
BibTeX
@article{Sharma2026,
author = {Sharma, Shweta},
title = {Graph Neural Network Based Modelling of Pest Spread Across Spatially Connected Agricultural Fields},
journal = {Journal of Smart and Sustainable Farming},
year = {2026},
pages = {062},
doi = {10.64026/jssf/2026007}
}
EUVIMED