Machine Learning Analysis of Habitat Disturbance and Parasite Loads in Wild Primates

Abstract

Wild primates serve an important function in ensuring ecosystem stability via seed dispersal, forest regeneration, and preservation of biodiversity. However, the increased habitat disturbance due to deforestation, land-use changes, human interaction, and environmental degradation have had considerable implications for the well-being and survival of primates. Habitat disturbance may alter the behavior of animals, make them more prone to infection-causing pathogens, and lead to the existence of different levels of parasites in wild primate communities. Traditional methods of parasite evaluation primarily rely on field sampling and laboratory analysis, which are tedious, involve experts' participation, and offer little capacity for ecological monitoring. This research introduces an ecological analysis tool using machine learning for evaluating habitat disturbance and parasite loads in wild primates. The suggested framework takes into consideration habitat quality parameters, vegetation parameters, human disturbance parameters, parasite infections, and primate health to detect ecological patterns related to parasite risks. Machine learning models, such as Random Forest, Support Vector Machine, Extreme Gradient Boosting, and Neural Network models, are used for the purpose of parasite load estimation and risk classification. As a result, it is shown that habitat fragmentation, less vegetation, and higher human disturbance parameters are key factors impacting the process of parasite infections and infections' loads. Thus, the suggested framework allows assessing the health state of the wild population in a data-driven manner and detecting high-risk habitats.