Presentation 2003/6/20
Modular Network SOM Self-Organizing Map of a Systems Group in Function Space
Kazuhiro TOKUNAGA, Tetsuo FURUKAWA, Syozo YASUI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Generalization of SOM called Modular Network SOM (MNSOM) is presented herein. In it, each unit of the conventional SOM is replaced by a functional module which is a multi-layer neural network. MNSOM learns a group of systems in terms of their input-output relationships, thus producing different models corresponding to the member systems. At the same time, each such model is allocated in a 2D-space according to the mutual nearness, on the basis of the Winner-Take- All policy as well as the neighborhood cooperation in much the same way as in the usual SOM. Thus, MNSOM is a SOM in the function space rather than the vector space, not merely allowing data classification in terms of the modular-network representation models, but also tells the distance between the models. The feasibility of MNSOM is demonstrated by an example relevant to a family of cubic functions as well as another involving a nation-wide meteorological dataset.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Modular Network / Self-Organizing Map / Neighborhood function
Paper # NC2003-17
Date of Issue

Conference Information
Committee NC
Conference Date 2003/6/20(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Modular Network SOM Self-Organizing Map of a Systems Group in Function Space
Sub Title (in English)
Keyword(1) Modular Network
Keyword(2) Self-Organizing Map
Keyword(3) Neighborhood function
1st Author's Name Kazuhiro TOKUNAGA
1st Author's Affiliation Graduate School of Life Science and System Engineering, Kyushu Institute of Technology()
2nd Author's Name Tetsuo FURUKAWA
2nd Author's Affiliation Faculty of Computer Science and Systems Engineering, Kyushu Institute of Technology
3rd Author's Name Syozo YASUI
3rd Author's Affiliation Graduate School of Life Science and System Engineering, Kyushu Institute of Technology
Date 2003/6/20
Paper # NC2003-17
Volume (vol) vol.103
Number (no) 153
Page pp.pp.-
#Pages 6
Date of Issue