Predicción de radiación solar mediante deep belief network
Fecha
Autor corporativo
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Distrital Francisco José de Caldas. Colombia
Compartir
Director
Altmetric
Resumen
Descripción
El desarrollo continuo de las herramientas computacionales ofrece la posibilidad de realizar procesos con la capacidad de llevar a cabo actividades con mayor eficiencia, exactitud y precisión. Entre estas herramientas se encuentra la arquitectura neuronal, Deep Belief Network (DBN), diseñada con el propósito de colaborar en el desarrollo de técnicas de predicción para hallar información que permita estudiar el comportamiento de los fenómenos naturales, como lo es la radiación solar. En el presente trabajo se presentan los resultados obtenidos al manejar la arquitectura DBN para predicción de radiación solar, la cual se simula mediante la herramienta de programación Visual Studio C#, indicando el nivel de profundidad que posee esta arquitectura, como afecta la cantidad de capas y de neuronas en el entrenamiento y los resultados obtenidos para poder predecir los valores deseados en el 2014, con errores cercanos al 2 % y mayor rapidez para el entrenamiento, respecto a errores obtenidos por métodos convencionales de entrenamiento neuronal, que se encuentran por el 5% y que a su vez llevan largos periodos de entrenamiento.
The continued development of computational tools offers the possibility to execute processes with the ability to carry out activities more efficiently, exactness and precision. Between these tools there is the neural architecture, Deep Belief Network (DBN), designed to collaborate in the development of prediction technics to find information that allows to study the behavior of the natural phenomena, such as the solar insolation. This paper presents the obtained results when using the DBN architecture for solar insolation prediction, simulated through the programming tool Visual Studio C#, showing the deep level that this architecture has, how it affects the number of layers and neurons per layer in the training and the results to predict the desired values in 2014, with errors close to 2% and faster to training, respect to errors obtained through conventional methods for neural training, which are about 5% and take long periods of training.
The continued development of computational tools offers the possibility to execute processes with the ability to carry out activities more efficiently, exactness and precision. Between these tools there is the neural architecture, Deep Belief Network (DBN), designed to collaborate in the development of prediction technics to find information that allows to study the behavior of the natural phenomena, such as the solar insolation. This paper presents the obtained results when using the DBN architecture for solar insolation prediction, simulated through the programming tool Visual Studio C#, showing the deep level that this architecture has, how it affects the number of layers and neurons per layer in the training and the results to predict the desired values in 2014, with errors close to 2% and faster to training, respect to errors obtained through conventional methods for neural training, which are about 5% and take long periods of training.
Palabras clave
Accord.Net, Afforge.Net, Back propagation (BP), Contrastive Divergence (CD), Deep Belief Network (DBN), Restricted Boltzmann Machine (RBM), solar insolation prediction, visual studio 2010 - C#., Accord.Net, Afforge.Net, Back propagation (BP), Contrastive Divergence (CD), Deep Belief Network (DBN), predicción de radiación solar, Restricted Boltzmann Machine (RBM), visual studio 2010 - C#.