Hierarchical Model Based On Generalized Additive Models Used To Estimate The Level Of PM₁₀ Concentration Effects On The Human Health

Objective:Outdoor air of pollution on human health will produce certain effect. Observational epidemiological studies from the perspective of population health provides the evidence about the health effects of air pollution. This article will focus on long-term daily monitoring data. Now based on the statistic analysis, we can make more exact inference.Methods:For a single city, the number of deaths in the PM10 impact on the residents of the estimates, using generalized additive models. Using smoothing spline to control the confounding factors such as time, temperature. When to estimate the level of PM10 concentration effects on the pepole's health in the total area, use hierarchical model. At the same time, implement the methods with a giving example.Results:In the six cities:det,chic,anch,denv,salt and spok, an increase of 10 ? g m3 of PM10 in yesterday corresponded to 0.071%, 0.053%, 0.005%, 0.083%, 0.008% and 0.028% respectively in six cities. After a general estimate, the area the day before the concentration of PM10 for every 10μg/m3, the day the number of deaths increased by 0.16%.Conclusion:When using hierarchical model based on generalized additive model to analyze the relation between air pollution and motality, the amount of information use full, the result is more fulfilling.This method is a useful way to resolve such issues.