Examinando por Autor "García Correa, David Eduardo"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemPrototipo de red neuronal artificial para el pronóstico de eventos críticos por partículas PM2.5 en el centro de la ciudad de Manizales.(Fundación Universitaria Los Libertadores. Sede Bogotá., ) García Correa, David Eduardo; Gonzalez Veloza, Jose John FredyThe following work aims to develop an air quality forecast model in the center of the city of Manizales - Caldas, starting from a more or less processed data set, it is desired to create a model that allows to successfully predict the behavior or value that takes the variable PM2.5 from 8 variables (Date, PM2.5, Temperature, Wind Speed, Precipitation, SO2, O3 and CO) and 603 data for each variable for a total of 264,299 data, between the years 2019 and 2020, in order to generate preventive measures for respiratory diseases that are diagnosed based on high levels of air pollution. The forecasting procedure was carried out by creating an artificial neural network, for which the data set was subdivided into train and test, in a subdivision of 567 consecutive days for network training and the next 30 days for its validation. . This is a proportion that was chosen because it is forecasting a month ahead, which is why it seemed convenient. Finally, this work presents the use of time series using Machine Learning techniques for the prediction of PM2.5, this tool and methodology had not been used until now for that purpose by the Environmental Authority - CORPOCALDAS. The results obtained allow to have confidence in the predictions of PM2.5 with an error of 3%, with which it is possible to have sufficient confidence to inform the community about the changes in the levels of air quality and concentration of PM2.5 in the near future