Self-Organizing Maps Applied to the Analysis of Atmospheric Pollutants
Machine Learning, Atmospheric Pollutants, Self-Organizing Maps, Meteorological Data, Salvador-BA.
This work aims to apply a new approach to data analysis of air pollutants through a machine learning technique. The technique is based on an unsupervised artificial neural network of the self-organizing maps type. The study was carried out in the city of Salvador - Bahia based on parameters obtained by air quality monitoring stations distributed along its length. Data were collected from 2011 to 2016 and made available by the Government of the State of Bahia through CETREL SA. An exploratory analysis was performed from the results achieved by applying self-organizing maps, showing correlations between pollutants and data weather, data groupings, possible emission sources, and conclusions about the problem.