Examinando por Autor "Reyes González, Alexander"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemPredicción de la velocidad del viento mediante un modelo GradientBoostingTree de machine learning aplicable en la gestión del tránsito aéreo(Fundación Universitaria Los Libertadores. Sede Bogotá., ) Reyes González, Alexander; González Veloza, José John FredyThrough this study, the Gradient Boosting Tree was identified as the best statistical model within a group of ten, which design was aimed at predicting wind intensity at different heights, for application in air traffic management, and particularly in the design of routes and flight procedures as activities related to airspace organization and management. Rawinsonde data recorded between 2015 and 2020 were analyzed, obtaining a mean square error value of 13.6 knots for all data corresponding to heights between 2,546 m (8,353 ft) and 36,978 m (106,442 ft). When applying the model to the data with a height limit of 15,240 m, the root mean square error was reduced to 8.78 knots; however, the results of the predictions showed a common tendency towards an approximate value of 15 knots, with which it was determined that they are not useful in practice in the presence of higher wind values. The solution to this problem was found in the generation of a new database from unique height records and the assignment of critical values (maximum and minimum) extracted from each predictor variable. Once the data were processed by the model, an RMSE of 9.24 knots was obtained. Finally, the model allowed the prediction of wind values for altitudes between 8,400 ft and 30,000 ft, with an interval of 100 ft and a total of 217 records. It was observed that the model is very useful in determining the probability of occurrence of the predicted data, thus facilitating the determination of a technical parameter required to consider that the value of wind speed is usable for flight procedures design purposes; that is, a probability of 2σ (95%) corresponding to a value of 43.05 knots.