Backpropagation Neural Network for the Prediction of PM10 Contamination Data

The prevention of respiratory diseases caused by high air pollution rates is an important issue in big cities, where industrialization and overpopulation cause an increase in allergenic particles that aggravate the disease of allergic rhinitis and asthma, especially in childhood. The problem lies in the disinformation of the population about air quality and the preventive measures to be taken in order to avoid deterioration in health. In this paper, data are monitored by a sensor network that registers the most abundant allergen, called PM10, for the city of León, Guanajuato. An artificial neural network (ANN) with a supervised Backpropagation training is used to predict future data until a minimum error is reached. The proposed methodology generates efficient results, measured in the error of the solutions and in execution time.