| Presentation | 2003/3/10 A New Kernel for Binary Regression Masashi SUGIYAMA, Maki FUJINO, Hidemitsu OGAWA, |
|---|---|
| PDF Download Page | PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | Recently, kernel machines have been extensively studied and shown to work surprisingly well on various real-world problems, given appropriate kernel functions. In this paper, we propose a new class of kernels called the principal component (PC) kernels for the binary regression problem. The PC kernel is designed based on the theory of the Karhunen-Loeve expansion, so it can approximate the binary learning target function very well. The PC kernels can effectively incorporate a prior knowledge of the learning target functions. In the absence of such a prior knowledge. we further derive a specific PC kernel called the cosine bell (CB) kernel using a non-informative prior knowledge. Computer simulations with ridge regression show that the CB kernel is comoared favorably to the conventional Gaussian kernel. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | kernel methods / binary regression / binary classification / Karhunen-Loeve expansion |
| Paper # | NC2002-150 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2003/3/10(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 | ENG |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | A New Kernel for Binary Regression |
| Sub Title (in English) | |
| Keyword(1) | kernel methods |
| Keyword(2) | binary regression |
| Keyword(3) | binary classification |
| Keyword(4) | Karhunen-Loeve expansion |
| 1st Author's Name | Masashi SUGIYAMA |
| 1st Author's Affiliation | Department of Computer Science, Tokyo Institute of Technology() |
| 2nd Author's Name | Maki FUJINO |
| 2nd Author's Affiliation | Department of Electorical and Electronic Engineering, Tokyo Institute of Technology |
| 3rd Author's Name | Hidemitsu OGAWA |
| 3rd Author's Affiliation | Department of Computer Science, Tokyo Institute of Technology |
| Date | 2003/3/10 |
| Paper # | NC2002-150 |
| Volume (vol) | vol.102 |
| Number (no) | 729 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |