Integrated structure selection and parameter optimisation for eng-genes neural models


Autoria(s): Connally, Patrick; Li, Kang; Irwin, George
Data(s)

01/08/2008

Resumo

The eng-genes concept involves the use of fundamental known system functions as activation functions in a neural model to create a 'grey-box' neural network. One of the main issues in eng-genes modelling is to produce a parsimonious model given a model construction criterion. The challenges are that (1) the eng-genes model in most cases is a heterogenous network consisting of more than one type of nonlinear basis functions, and each basis function may have different set of parameters to be optimised; (2) the number of hidden nodes has to be chosen based on a model selection criterion. This is a mixed integer hard problem and this paper investigates the use of a forward selection algorithm to optimise both the network structure and the parameters of the system-derived activation functions. Results are included from case studies performed on a simulated continuously stirred tank reactor process, and using actual data from a pH neutralisation plant. The resulting eng-genes networks demonstrate superior simulation performance and transparency over a range of network sizes when compared to conventional neural models. (c) 2007 Elsevier B.V. All rights reserved.

Identificador

http://pure.qub.ac.uk/portal/en/publications/integrated-structure-selection-and-parameter-optimisation-for-enggenes-neural-models(868d39b8-f817-4c78-8a94-f4158cf0a389).html

http://dx.doi.org/10.1016/j.neucom.2007.06.005

http://www.scopus.com/inward/record.url?scp=56449089569&partnerID=8YFLogxK

Idioma(s)

eng

Direitos

info:eu-repo/semantics/restrictedAccess

Fonte

Connally , P , Li , K & Irwin , G 2008 , ' Integrated structure selection and parameter optimisation for eng-genes neural models ' Neurocomputing , vol 71 , no. 13-15 , pp. 2964-2977 . DOI: 10.1016/j.neucom.2007.06.005

Palavras-Chave #/dk/atira/pure/subjectarea/asjc/1700/1702 #Artificial Intelligence #/dk/atira/pure/subjectarea/asjc/1700/1706 #Computer Science Applications #/dk/atira/pure/subjectarea/asjc/2800/2805 #Cognitive Neuroscience
Tipo

article