| Presentation | 1996/2/3 Adaptive Constitution Method of Hidden Layer in Neural Network with RBFs Kazuko ISHIDA, Keisuke KAMEYAMA, Yukio KOSUGI, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | In multilayered neural networks homogeneous hidden layers exclusively consisting of sigmoidal functions or RBFs (Radial Basis Functions) are generally used. They discreminate classes with hyperplanes or hyperellipses respectivly in clustering problems. We consider a heterogeneous hidden layer consisting of these different charactaristic functions to classify patterns unknown distributed. Incremental auto-constitution algorithm and a rule which replaces RBF unit with a set of sigmoidal units are applied to look for the optimum constituent ratio of hidden units dynamically, then hidden units are pruned to improve generalization ability of tne network. The result showed that the proposed network is applicable not only for simple sorting problems but also for classification of complicatedly discribed patterns and for function approximation problems. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | RBF / clustering / hidden layer size / pruning / function approximation |
| Paper # | NC95-112 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 1996/2/3(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) | Adaptive Constitution Method of Hidden Layer in Neural Network with RBFs |
| Sub Title (in English) | |
| Keyword(1) | RBF |
| Keyword(2) | clustering |
| Keyword(3) | hidden layer size |
| Keyword(4) | pruning |
| Keyword(5) | function approximation |
| 1st Author's Name | Kazuko ISHIDA |
| 1st Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering Tokyo Institute of Technology() |
| 2nd Author's Name | Keisuke KAMEYAMA |
| 2nd Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering Tokyo Institute of Technology |
| 3rd Author's Name | Yukio KOSUGI |
| 3rd Author's Affiliation | Interdisciplinary Graduate School of Science and Engineering Tokyo Institute of Technology |
| Date | 1996/2/3 |
| Paper # | NC95-112 |
| Volume (vol) | vol.95 |
| Number (no) | 506 |
| Page | pp.pp.- |
| #Pages | 8 |
| Date of Issue | |