• English 
  • Spanish 

Applications of Bayesian Networks in Meteorology

Initial page: 
309
Last page: 
327

PDF file

In this paper we present some applications of Bayesian networks in Meteorology from a data mining point of view. We work with a database of observations (daily rainfall and maximum wind speed) in a network of 100 stations in the Iberian peninsula and with the corresponding gridded atmospheric patterns generated by a numerical circulation model. As a first step, we analyze the efficiency of standard learning algorithms to obtain directed acyclic graphs representing the spatial dependencies among the variables included in the database; we also present a new local learning algorithm which takes advantage of the spatial character of the problem. The resulting graphical models are applied to different meteorological problems including weather forecast and stochastic weather generation. Some promising results are reported.

Complete book details
2004_ABN_Portada.jpg
Book title: 
Advances in Bayesian Networks
Authors/Editors: 
Gámez, J.A.
Authors/Editors: 
Moral, S.
Authors/Editors: 
Salmerón, A.
Editorial: 
Sringer Verlag
City: 
Berlin
Year: 
2004
ISBN: 
978-3-540-20876-1