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

Area of Science:

• Environmental Science
• Epidemiology
• Data Science

Background:

• Dengue fever is transmitted by Aedes mosquitoes, necessitating effective population control.
• Traditional gravitrap monitoring often underestimates Aedes mosquito populations and lacks spatial-temporal detail.
• Limited data exists on urban Aedes population dynamics, hindering targeted vector control efforts.

Purpose of the Study:

• To develop a novel index for assessing spatial-temporal dynamics of adult Aedes mosquitoes in urban environments.
• To improve the accuracy and efficiency of Aedes mosquito surveillance compared to traditional methods.
• To provide a tool for guiding public health interventions against dengue fever.

Main Methods:

• An artificial intelligence (AI) surveillance system utilizing an auto-Markov model with a non-parametric permutation test was developed.
• The auto-Markov model incorporates neighborhood effects to dynamically adjust spatial-temporal risks based on environmental factors.
• Information from adjacent villages was integrated to enhance the precision of Aedes population risk prediction.

Main Results:

• The AI gravitrap index demonstrated enhanced sensitivity in predicting Aedes densities by integrating auto-Markov and disease mapping models.
• Simulation and cross-validation studies confirmed the AI index's superior efficiency over traditional indices in risk assessment.
• The AI index can reduce the cost of gravitrap deployment and provides more accurate spatial-temporal risk maps.

Conclusions:

• The AI gravitrap index offers a flexible and dynamic tool for updating Aedes mosquito risk levels, applicable across diverse urban settings.
• The index's ability to accommodate spatial-temporal dependencies ensures more accurate reflection of vector population dynamics.
• AI-driven risk maps can effectively guide policymakers in preventing dengue epidemics.