Forecast of Air Quality Based on Ozone by Decision Trees and Neural Networks

In this paper we explore models based on decision trees and neural networks models for predicting levels of ozone. We worked with a data set of the Atmospheric Monitoring System of Mexico City (SIMAT), which includes measurements hour by hour, between 2010 to 2011. The data come from of three meteorological stations: Pedregal, Tlalnepantla and Xalostoc in Mexico city. The data set includes 8 parameters: four chemical variables and four meteorological variables. Based on our results, it's possible to predict ozone levels with these parameters, with an accuracy of 94.4%.