Presentation | 2001/6/22 Generalization Performence : SVM vs Complexity-Regularization Naoki Tsukamoto, Haruhisa Takahashi, |
---|---|
PDF Download Page | PDF download Page Link |
Abstract(in Japanese) | (See Japanese page) |
Abstract(in English) | Support vector machine is paid great attention recently from machine learning and neural network societies. This method is being a standard method for machine learning, which well meets trade-off for quick learning and good generalization. In this report we compare on the generalization performance the SVM with the complexity regularization which minimizes the similar loss function as SVM. The result on the Majority XOR problem says that if input dimension is small the complexity regularization shows better performance and vice versa. The experiment results also says that if the data size is large compared with the input dimension complexity regularization is better than SVM. |
Keyword(in Japanese) | (See Japanese page) |
Keyword(in English) | Support vectoer machine / Complexity-Regularization |
Paper # | NC2001-29 |
Date of Issue |
Conference Information | |
Committee | NC |
---|---|
Conference Date | 2001/6/22(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) | Generalization Performence : SVM vs Complexity-Regularization |
Sub Title (in English) | |
Keyword(1) | Support vectoer machine |
Keyword(2) | Complexity-Regularization |
1st Author's Name | Naoki Tsukamoto |
1st Author's Affiliation | Department of Communications and Systems, The University of Electro-Communications() |
2nd Author's Name | Haruhisa Takahashi |
2nd Author's Affiliation | Department of Communications and Systems, The University of Electro-Communications |
Date | 2001/6/22 |
Paper # | NC2001-29 |
Volume (vol) | vol.101 |
Number (no) | 154 |
Page | pp.pp.- |
#Pages | 7 |
Date of Issue |