Garcia-Laencina P.J.,Centro Universitario Of La Defensa Of San Javier Mde Upct |
Roca-Gonzalez J.-L.,Centro Universitario Of La Defensa Of San Javier Mde Upct |
Bueno-Crespo A.,San Antonio de Murcia Catholic University |
Sancho-Gomez J.-L.,Technical University of Cartagena
Neural Network World | Year: 2013
Extreme learning machine (ELM) is an emergent method for training single hidden layer feedforward neural networks (SLFNs) with extremely fast training speed, easy implementation and good generalization performance. This work presents effective ensemble procedures for combining ELMs by exploiting diversity. A large number of ELMs are initially trained in three different scenarios: the original feature input space, the obtained feature subset by forward selection and different random subsets of features. The best combination of ELMs is constructed according to an exact ranking of the trained models and the useless networks are discarded. The experimental results on several regression problems show that robust ensemble approaches that exploit diversity can effectively improve the performance compared with the standard ELM algorithm and other recent ELM extensions. © CTU FTS 2013.
Oton-Martinez R.A.,EXPAL Systems S.A. |
Monreal-Gonzalez G.,Technical University of Cartagena |
Garcia-Cascales J.R.,Technical University of Cartagena |
Vera-Garcia F.,Technical University of Cartagena |
And 2 more authors.
International Journal for Numerical Methods in Fluids | Year: 2015
A model formulated in terms of conserved variables is proposed for its use in the study of internal ballistic problems of pyrotechnical mixtures and propellants. It is a transient two-phase flow model adapted from the non-conservative Gough model. This conversion is mathematically attractive because of the wide range of numerical methods for this kind of systems that may be applied. We propose the use of the AUSM+, AUSM+up and Rusanov schemes as an efficient alternative for this type of two-phase problem. A splitting technique is applied, which solves the system of equations in several steps. A second-order approach based on Monotonic Upstream-Centred Scheme for Conservation Laws (MUSCL) is also used. Some tests are used to validate the code, namely a shock wave test, a contact discontinuity problem and an internal ballistics problem. In this last case, one-dimensional numerical results are compared with experimental data of 155-mm gunshots. © 2015 John Wiley & Sons, Ltd.