Projected typical allergic diseases prevalence under changing environments based on multiple machine learning models

Timely understanding the prevalence of allergic skin diseases (ASD) and allergic nasopharyngeal disease (AND) is essential for effective public health planning and resource allocation. However, accurately predicting ASD and AND poses a significant challenge due to the complex interplay of environmental and individual factors. A machine learning-based scheme was proposed for predicting the prevalence of ASD and AND using environmental and hydrological data (n = 85). Significant variations in predictive accuracy were observed across different algorithms. For ASD, the decision tree regression (DTR) demonstrated the best performance. For AND, the ridge regression (RR) model yielded the best results, respectively. Based on Urumqi's 2022 population, the projected peak number of individuals with ASD is expected to rise by 215,000, 243,200, and 275,600 compared to January 2015. For AND, the projected peak increases are expected to be 38,900, 35,700, and 56,300, respectively. Environmental factors exhibit significant correlations with the prevalence of ASD and AND, with minimum temperature identified as the most influential factor affecting both conditions. Machine learning models that incorporate these environmental variables were proven to effectively predict the prevalence of both conditions. Based on the model's projections under three climate change scenarios, a significant increase in the prevalence of ASD and AND in Urumqi is expected from 2015 to 2099. This trend underscores the potential impact of climate change on public health in the region, highlighting the need for proactive measures to address these emerging challenges.