Model-based neural networks

Terry M. Caelli, David Mc G. Squire, Tom P.J. Wild

Research output: Contribution to journalArticleResearchpeer-review

27 Citations (Scopus)


In this paper we show how neural networks can be formulated in terms of various parameterised connection models which explicitly encode desired properties of the target system. Such a modelling approach to neural networks raises issues about their relationships to other technologies such as Adaptive Filtering and Principal Components Analysis. The benefits of this approach can be a significant decrease in the parameter space, improved generalisation, and a learning procedure which guarantees a priori specified invariance constraints.

Original languageEnglish
Pages (from-to)613-625
Number of pages13
JournalNeural Networks
Issue number5
Publication statusPublished - 1 Jan 1993
Externally publishedYes


  • Adaptive filters
  • Classification
  • Invariance
  • Number of parameters

Cite this