RT info:eu-repo/semantics/article
T1 Multivariate influence through neural networks ensemble: Study of Saharan dust intrusion in the Canary Islands
A1 González Mendoza, Luis Antonio
A1 Gonzalez-Calvo, D.
A1 Aguilar, R.M.
A1 Criado-Hernández, C.
A2 Ingeniería Química y Tecnología Farmacéutica
K1 Neural networks
K1 Ensemble methods
K1 Relative importance
K1 Calima
K1 Saharan dust
K1 Canary Islands
AB Analyzing and predicting the concentration of airborne dust is vital to the economic activity and tothe health of the population. In this study, we use a set of artificial neural networks that we structurethrough ensemble learning to yield a complex variable, such as the concentration of dust, based onactual data such as air temperature, relative humidity, atmospheric pressure and wind speed. Thestatistical performance indices obtained, show the effectiveness of the proposed approach through theapplication of a cross-validation committee. It is thus vital to have a reliable calculation method fordetermining relative importances that can be applied to this type of ensemble architecture by way ofartificial neural networks.Unlike other relative importance methods, where calculations are done based directly on theweights in the artificial neural network and whose results in ensemble sets exhibit high dispersion,we propose our own procedure, which selectively chooses the variation in the inputs to readjust thearchitecture of the neural network. This allows us to measure those variables with the greatest effecton the target variable, thus obtaining the multivariate influence on the surface dust concentrationthrough a computational model.This method thus provides a real alternative for calculating and estimating relative importance thatcan be generalized to any type of problem for multivariate systems modeled using artificial neuralnetworks for both, a simple configuration, and an ensemble architecture
YR 2021
FD 2021
LK http://riull.ull.es/xmlui/handle/915/34849
UL http://riull.ull.es/xmlui/handle/915/34849
LA en
DS Repositorio institucional de la Universidad de La Laguna
RD 28-nov-2024