Controlling and looking for possible sources of air pollution has become a relevant goal in our increasingly degraded environment. In this paper the authors set up the fundamentals to propose, build and further develop a model to control and measure air pollution. The model is based on modern applied statistical techniques. Particularly, locally weighted nonparametric regression procedures proved to better fit the empirical data giving a more detailed information concerning the pollution development. The dependence on past events has been determined by means of dynamic time series models. The area of study has been selected to represent a mediterranean area, so that the conclusions can be easily generalised to mediterranean polluted environments. (A)
Abstract