Artificial Intelligence Analysis of Rodent Communities for Zoonotic Risk Surveillance

Abstract

Rodents are a key part of terrestrial ecosystems, playing vital roles in food web interactions, seed dispersal, and nutrient cycling. They are important natural reservoirs of zoonotic pathogens, however, and thus pose a concern with regard to wildlife health and the emergence of diseases. The dynamics of the populations can be affected by changes in habitat, climate variability, and by the rate of human–wildlife interaction, which can increase the likelihood of pathogen transmission. Traditional rodent monitoring techniques primarily rely on field surveys and trapping and offer poor analytical capacity for complex ecological relationships and are time consuming and resource intensive. This study introduces an approach to analyse rodent community using artificial intelligence (AI) to aid zoonotic risk surveillance. The suggested model combines ecological, behavioural and environmental factors (abundance of rodents, habitat characteristics, vegetation status, climatic parameters, anthropogenic influence). Patterns related to rodent distribution are detected with machine learning models and potential zoonotic risk conditions are predicted. The results reveal that the relationships between the environmental changes and the changes in the rodent communities can be successfully captured by the AI based models and the Random Forest model gives a better prediction performance than other models. Four key factors were determined to influence zoonotic risk patterns: habitat disturbance, temperature, availability of vegetation, and proximity to humans. The results suggest that AI can be used as a decision-support tool for wildlife disease monitoring, as it can help identify areas with high risk for early detection. This study can help advance One Health approaches for surveillance that aim to combine ecological, wildlife, and zoonotic disease surveillance and control