Monolithic neural networks may be trained from measured data to establish knowledge about the process. Unfortunately, this knowledge is not guaranteed to be found and – if at all – hard to extract. Modular neural networks are better suited for this purpose. Domain-ordered by topology, rule extraction is performed module by module. This has all the benefits of a divide-and-conquer method and opens the way to structured design. This paper discusses a next step in this direction by illustrating the potential of base functions to design the neural model.
|Title of host publication||Proceedings of the Fourteenth Belgium/Netherlands Conference on Artificial Intelligence (BNAIC'02)|
|Editors||Hendrik Blockeel, Marc Denecker|
|Place of Publication||Leuven, Belgium|
|Number of pages||2|
|Publication status||Published - Oct 2002|
|Event||14th Belgium-Dutch Conference on Artificial Intelligence, BNAIC 2002 - Leuven, Belgium|
Duration: 21 Oct 2002 → 22 Oct 2002
Conference number: 14
|Conference||14th Belgium-Dutch Conference on Artificial Intelligence, BNAIC 2002|
|Period||21/10/02 → 22/10/02|
van der Zwaag, B. J., Slump, C. H., & Spaanenburg, L. (2002). Process identification through modular neural networks and rule extraction (extended abstract). In H. Blockeel, & M. Denecker (Eds.), Proceedings of the Fourteenth Belgium/Netherlands Conference on Artificial Intelligence (BNAIC'02) (pp. 507-508). Leuven, Belgium.