| Presentation | 2012-01-27 A statistical analysis of soft-margin support vector machines for non-separable problems Hiroyuki FUNAYA, Kazushi IKEDA, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The statistical properties of SVMs for non-separable problems are studied. SVMs with hard margins are not always solvable for non-separable problems. Introducing soft margin alleviates this difficulty, but an SVM still fails in successfully solving the problems for heavily overlapped data. In this study, the probability that an SVM solves a problem properly is mathematically derived in one-dimensional case for both hard-margin's and soft-margin's case. Some computer simulations confirm the theoretical validity. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | SVM / ν-SVM / adaptive / statistical analysis |
| Paper # | NC2011-109 |
| Date of Issue |
| Conference Information | |
| Committee | NC |
|---|---|
| Conference Date | 2012/1/19(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 statistical analysis of soft-margin support vector machines for non-separable problems |
| Sub Title (in English) | |
| Keyword(1) | SVM |
| Keyword(2) | ν-SVM |
| Keyword(3) | adaptive |
| Keyword(4) | statistical analysis |
| 1st Author's Name | Hiroyuki FUNAYA |
| 1st Author's Affiliation | Nara Institute of Science and Technology Ikoma() |
| 2nd Author's Name | Kazushi IKEDA |
| 2nd Author's Affiliation | Nara Institute of Science and Technology Ikoma |
| Date | 2012-01-27 |
| Paper # | NC2011-109 |
| Volume (vol) | vol.111 |
| Number (no) | 419 |
| Page | pp.pp.- |
| #Pages | 6 |
| Date of Issue | |