An improved particle swarm optimization (PSO): method to enhance modeling of airborne particulate matter (PM10)

Nowadays, it is of paramount importance for human health the monitoring and modelling of air quality. Among the different pollutants, there are some that are considerable more difficult to model due to their chemical composition. Some of these are particulate matter (particles ≤ 10 microns, PM10, and particles ≤ 2.5 microns, PM2.5), which can cause respiratory diseases or even cause premature deaths. Furthermore, There are several models that can be used to evaluate air quality. In this contribution, the combination of a neuro-fuzzy based method with particle swarm optimization is proposed to crucially increase accuracy when dealing with the non-linear behavior of airborne particulate matter (PM10). Several experiments were carried out to show the feasibility of the proposed method and to show that even when the nature of the data, that has dynamic behavior, variance in the spread of data, the present of outliers, climatic conditions, among other factors, the modeling of this particular set of data may be made accurately, showing the robustness and feasibility of the proposed method.