Patchwork neuro-fuzzy system with hierarchical domain partition

Simiński, Krzysztof

doi:10.1007/978-3-540-93905-4_2

Krzysztof Simiński

Abstract:

The paper presents the patchwork hierarchical domain partition in the neuro-fuzzy system with parameterized consequences. The hierarchical domain par- tition has the advantages of grid partition and clustering. It avoids the curse of dimensionality and reduces the occurrence of areas with low membership to all re- gions. The paper depicts the iterative hybrid procedure of hierarchical split. The splitting procedure estimates the best way of creating of the new region: (1) based on ﬁnding and splitting the region with the highest contribution to the error of the system or (2) creation of patch region for the highest error area. The paper presents the results of experiments on real life and synthetic datasets. This approach can produce neuro-fuzzy inference systems with better generalisation ability and subse- quently lower error rate.

Reference:

Krzysztof Simiński, Patchwork neuro-fuzzy system with hierarchical domain partition, [chapter in] Computer Recognition Systems 3 (Marek Kurzyński, Michał Woźniak, eds.), Springer-Verlag, 2009, volume 57, pp. 11-18. ([10 pkt])

Bibtex Entry:

@INCOLLECTION{id:Siminski2009Patchwork,
  author = {Krzysztof Simi\'{n}ski},
  title = {Patchwork neuro-fuzzy system with hierarchical domain partition},
  booktitle = {Computer Recognition Systems 3},
  publisher = {Springer-Verlag},
  year = {2009},
  editor = {Marek Kurzy\'{n}ski and Micha{\l} Wo\'{z}niak},
  volume = {57},
  series = {Advances in Intelligent and Soft Computing},
  pages = {11--18},
  address = {Berlin, Heidelberg},
  timestamp = {2009.02.26},
  doi = {10.1007/978-3-540-93905-4_2},
  abstract = {The paper presents the patchwork hierarchical domain partition in the
neuro-fuzzy system with parameterized consequences. The hierarchical domain par-
tition has the advantages of grid partition and clustering. It avoids the curse of
dimensionality and reduces the occurrence of areas with low membership to all re-
gions. The paper depicts the iterative hybrid procedure of hierarchical split. The
splitting procedure estimates the best way of creating of the new region: (1) based
on ﬁnding and splitting the region with the highest contribution to the error of the
system or (2) creation of patch region for the highest error area. The paper presents
the results of experiments on real life and synthetic datasets. This approach can
produce neuro-fuzzy inference systems with better generalisation ability and subse-
quently lower error rate.},
note = {[10 pkt]},

}