An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk

Abstract Background Dengue fever is transmitted to humans through bites of Aedes mosquito vectors. Therefore, controlling the Aedes population can decrease the incidence and block transmission of dengue fever and other diseases transmitted by these mosquito species. In many countries, gravitraps are...

Full description

Saved in:
Bibliographic Details
Main Authors: Hsiang-Yu Yuan, Pei-Sheng Lin, Wei-Liang Liu, Tzai-Hung Wen, Yu-Chun Lu, Chun-Hong Chen, Li‑Wei Chen
Format: Article
Language:English
Published: BMC 2025-08-01
Series:International Journal of Health Geographics
Subjects:
Online Access:https://doi.org/10.1186/s12942-025-00403-z
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849761334297624576
author Hsiang-Yu Yuan
Pei-Sheng Lin
Wei-Liang Liu
Tzai-Hung Wen
Yu-Chun Lu
Chun-Hong Chen
Li‑Wei Chen
author_facet Hsiang-Yu Yuan
Pei-Sheng Lin
Wei-Liang Liu
Tzai-Hung Wen
Yu-Chun Lu
Chun-Hong Chen
Li‑Wei Chen
author_sort Hsiang-Yu Yuan
collection DOAJ
description Abstract Background Dengue fever is transmitted to humans through bites of Aedes mosquito vectors. Therefore, controlling the Aedes population can decrease the incidence and block transmission of dengue fever and other diseases transmitted by these mosquito species. In many countries, gravitraps are used to monitor mosquito vector densities, but this approach usually underestimates the population of Aedes mosquitoes. Moreover, literature on the spatio-temporal dynamics of Aedes populations in a single city is limited. For example, in Kaohsiung of Taiwan, population densities vary substantially between villages, and the city government has relatively limited resources to deploy gravitraps. Therefore, a well-defined index should be developed to reflect the spatial–temporal dynamics of adult Aedes mosquitoes in urban environments. This would allow reduction of sources and removal of vector habitats under various situations. Methods An artificial intelligence (AI) surveillance based on an auto-Markov model with a non-parametric permutation test is proposed. The auto-Markov model takes neighborhood effects into consideration, and can therefore adjust spatial–temporal risks dynamically in various seasons and environmental background. Information from neighboring villages is incorporated into the model to enhance precision of risk prediction. Results The proposed AI gravitrap index integrates the auto-Markov and disease mapping models to enhance sensitivity in risk prediction for Aedes densities. Simulation studies and cross-validation analysis indicated that the AI index could be more efficient than traditional indices in assessing risk levels. This means that using the AI index could also reduce allocation cost for gravitraps. Moreover, since the auto-Markov model accommodates spatial–temporal dependence, a risk map by the AI index could reflect spatial–temporal dynamics for Aedes densities more accurate. Conclusions The AI gravitrap index can dynamically update risk levels by the auto-Markov model with an unsupervised permutation test. The proposed index thus has flexibility to apply in various cities with different environmental background and weather conditions. Furthermore, a risk map by the AI index could provide guidance for policymakers to prevent dengue epidemics.
format Article
id doaj-art-88b0cbc85e49404dbc82eadd990277f9
institution DOAJ
issn 1476-072X
language English
publishDate 2025-08-01
publisher BMC
record_format Article
series International Journal of Health Geographics
spelling doaj-art-88b0cbc85e49404dbc82eadd990277f92025-08-20T03:06:04ZengBMCInternational Journal of Health Geographics1476-072X2025-08-0124111410.1186/s12942-025-00403-zAn AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes riskHsiang-Yu Yuan0Pei-Sheng Lin1Wei-Liang Liu2Tzai-Hung Wen3Yu-Chun Lu4Chun-Hong Chen5Li‑Wei Chen6Department of Biomedical Sciences, City University of Hong Kong, College of BiomedicineInstitute of Population Health Sciences, National Health Research InstitutesNational Mosquito-Borne Diseases Control Research Center, National Health Research InstitutesDepartment of Geography, National Taiwan UniversityInstitute of Population Health Sciences, National Health Research InstitutesNational Mosquito-Borne Diseases Control Research Center, National Health Research InstitutesInstitute of Population Health Sciences, National Health Research InstitutesAbstract Background Dengue fever is transmitted to humans through bites of Aedes mosquito vectors. Therefore, controlling the Aedes population can decrease the incidence and block transmission of dengue fever and other diseases transmitted by these mosquito species. In many countries, gravitraps are used to monitor mosquito vector densities, but this approach usually underestimates the population of Aedes mosquitoes. Moreover, literature on the spatio-temporal dynamics of Aedes populations in a single city is limited. For example, in Kaohsiung of Taiwan, population densities vary substantially between villages, and the city government has relatively limited resources to deploy gravitraps. Therefore, a well-defined index should be developed to reflect the spatial–temporal dynamics of adult Aedes mosquitoes in urban environments. This would allow reduction of sources and removal of vector habitats under various situations. Methods An artificial intelligence (AI) surveillance based on an auto-Markov model with a non-parametric permutation test is proposed. The auto-Markov model takes neighborhood effects into consideration, and can therefore adjust spatial–temporal risks dynamically in various seasons and environmental background. Information from neighboring villages is incorporated into the model to enhance precision of risk prediction. Results The proposed AI gravitrap index integrates the auto-Markov and disease mapping models to enhance sensitivity in risk prediction for Aedes densities. Simulation studies and cross-validation analysis indicated that the AI index could be more efficient than traditional indices in assessing risk levels. This means that using the AI index could also reduce allocation cost for gravitraps. Moreover, since the auto-Markov model accommodates spatial–temporal dependence, a risk map by the AI index could reflect spatial–temporal dynamics for Aedes densities more accurate. Conclusions The AI gravitrap index can dynamically update risk levels by the auto-Markov model with an unsupervised permutation test. The proposed index thus has flexibility to apply in various cities with different environmental background and weather conditions. Furthermore, a risk map by the AI index could provide guidance for policymakers to prevent dengue epidemics.https://doi.org/10.1186/s12942-025-00403-zAedes indexAI methodAuto-Markov modelDengue preventionGravitrapSpatial–temporal patterns
spellingShingle Hsiang-Yu Yuan
Pei-Sheng Lin
Wei-Liang Liu
Tzai-Hung Wen
Yu-Chun Lu
Chun-Hong Chen
Li‑Wei Chen
An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
International Journal of Health Geographics
Aedes index
AI method
Auto-Markov model
Dengue prevention
Gravitrap
Spatial–temporal patterns
title An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
title_full An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
title_fullStr An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
title_full_unstemmed An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
title_short An AI-based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
title_sort ai based gravitrap surveillance for spatial interaction analysis in predicting aedes risk
topic Aedes index
AI method
Auto-Markov model
Dengue prevention
Gravitrap
Spatial–temporal patterns
url https://doi.org/10.1186/s12942-025-00403-z
work_keys_str_mv AT hsiangyuyuan anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT peishenglin anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT weiliangliu anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT tzaihungwen anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT yuchunlu anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT chunhongchen anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT liweichen anaibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT hsiangyuyuan aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT peishenglin aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT weiliangliu aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT tzaihungwen aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT yuchunlu aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT chunhongchen aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk
AT liweichen aibasedgravitrapsurveillanceforspatialinteractionanalysisinpredictingaedesrisk