Presentation | 2005/10/11 Fisher Discriminant with Asymmetric Kernel Naoya KOIDE, Yukihiko YAMASHITA, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Although the progress of computer with respect to calculation and communication speeds is very rapid, the ability for cognitive processing such as pattern recognition is inferior to that of men. In this research, we apply the asymmetric kernel method to Fisher's linear discriminant and propose the variable kernel Fisher discriminant which can change the kernel parameter according to the position. In the learning stage, the parameter of boundary complexity for the learning sample is optimized. On the other hand the parameter for the unknown input pattern is fixed. Then, the proposed method can manage the case that the discriminant boundary is complex locally. In this paper, we provide the calculation method for Fisher discriminant with the asymmetric kernel and experimental results. We show the advantage of the method by experimental results where the method outperforms KFD for many datasets. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | kernel method / asymmetric kernel method / support vector machine (SVM) / kenrel Fisher discriminant |
Paper # | NC2005-48 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 2005/10/11(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) | Fisher Discriminant with Asymmetric Kernel |
Sub Title (in English) | |
Keyword(1) | kernel method |
Keyword(2) | asymmetric kernel method |
Keyword(3) | support vector machine (SVM) |
Keyword(4) | kenrel Fisher discriminant |
1st Author's Name | Naoya KOIDE |
1st Author's Affiliation | Graduate School of Sience and Engineering, Tokyo Institute of Technology() |
2nd Author's Name | Yukihiko YAMASHITA |
2nd Author's Affiliation | Graduate School of Sience and Engineering, Tokyo Institute of Technology |
Date | 2005/10/11 |
Paper # | NC2005-48 |
Volume (vol) | vol.105 |
Number (no) | 342 |
Page | pp.pp.- |
#Pages | 6 |
Date of Issue |