Nonlinear analysis of factors influencing public health in China based on machine learning

Public health is a key component of the United Nations Sustainable Development Goals (SDGs) and is central to the all-round development of individuals. Based on panel data from 31 provinces in China between 2010 and 2023, this article adopts the Population Mortality Rate (PMR) as a measure of public health and employs the Generalized Additive Model (GAM) from machine learning to systematically investigate the nonlinear effects of multidimensional factors-including medical resources, environmental pollution, socioeconomic conditions, and technology-on public health. The results of the study show that China's PMR exhibits an overall upward trend, with a spatially uneven distribution and significant regional disparities, as mortality rates in the central and western regions are generally higher than those in the eastern coastal provinces. All influencing factors show significant associations with the PMR. The effects of all influencing factors on mortality exhibit complex nonlinear characteristics, with their impacts varying considerably across different value ranges. Specifically, the influence of medical resources exhibits critical thresholds, such as the Number of Urban Practicing (assistant) Physicians per 10,000 people (NUPP) reaching its maximum health benefit at approximately 40; while the relationship between environmental indicators and mortality reveals potential complex confounding mechanisms, as evidenced by the negative association observed between Sulphur Dioxide Emissions (SDE) and mortality rates within the observed range; whereas the health benefits of basic resources and developmental factors, such as Per Capita Water Resources (PCWR), stabilize after crossing a specific threshold (20,000 cubic meters per person). Finally, some practical policy recommendations are put forward.