A procedure for the selection of neural models of dynamical processes is presented. It uses statistical tests at various levels of model reduction, in order to provide optimal tradeoffs between accuracy and parsimony. The efficiency of the method is illustrated by the modeling of a highly non-linear NARX process.

Chaotic time-series is a dynamic nonlinear system whose features can not be fully reflected by Linear Regression Model or Static Neural Network. While Nonlinear Autoregressive with eXogenous input includes feedback of network output, therefore, it can better reflect the system’s dynamic feature. Take annual active times of sunspot as an example, after verifying the chaos of sunspot time-series ...

a Dipartimento di Scienza Applicata e Tecnologia, Politecnico di Torino, Torino, 10129, Italy b NanoFacility Piemonte, Istituto Nazionale di Ricerca Metrologica (INRIM), 10135, Torino. c Optics & Photonics Technology Laboratory, Ecole Polytechnique Fédérale de Lausanne (EPFL), rue de la Maladière 71b, Neuchâtel CH-2002, Switzerland. d Fraunhofer Institute for Applied Optics and Precision Engine...

In this paper, a new correlation analysis based algorithm is proposed for the identification of a class of nonlinear systems which can be described by the NARX (Nonlinear AutoRegressive with eXogeneous input) model with input nonlinearities. Without any assumptions about the structure of an approximating function for the system nonlineariy, the algorithm recovers the functional values of the no...

The International Society for Clinical Densitometry (ISCD) and the International Osteoporosis Foundation (IOF) convened the FRAX(®) Position Development Conference (PDC) in Bucharest, Romania, on November 14, 2010, following a two-day joint meeting of the ISCD and IOF on the "Interpretation and Use of FRAX(®) in Clinical Practice." These three days of critical discussion and debate, led by a pa...

This paper describes an approach to predict non-verbal cues from speech-related features. Our previous investigations of audiovisual speech showed that there are strong correlations between the two modalities. In this work we developed two models using different kinds of Recurrent Artificial Neural Networks: Elman and NARX, to predict parameters of activity for head motion using linguistic and ...

BACKGROUND There is no well-defined means to identify the level of oral feeding skills (OFS) in preterm infants. OBJECTIVE To determine whether OFS as reflected by the combination of proficiency (PRO, %ml taken during the first 5 min/ml prescribed) and rate of milk transfer (RT, ml/min) correlates with gestational age (GA), oral feeding performance (OT, %ml taken during a feeding/ml prescribe...

Bill G. Horne NEC Research Institute 4 Independence Way Princeton, NJ 08540 c. Lee Gilest NEC Research Institute 4 Independence Way Princeton, N J 08540 It has recently been shown that gradient descent learning algorithms for recurrent neural networks can perform poorly on tasks that involve long-term dependencies. In this paper we explore this problem for a class of architectures called NARX n...

Agricultural cooperatives around the world are facing major structural challenges as they respond to a more industrialized agriculture, globalization, and freer trade. Co-ops are responding to these changes by merging, by finding new ways of raising capital, and by developing new organizational forms such as New Generation Co-ops. One of the critically important issues co-ops face as they under...

Nitric oxide (NO) is a highly reactive signaling molecule and plays a variety of physiological roles in plants. The research on the application of NO to postharvest preservation of flowers and fruits shows great promise in recent years. However, the physiological mechanism of exogenous NO to affect cut flowers is not very clear, and NO donor treatment protected plants from damage by increasing ...