| Presentation | 2021-03-02 Kernel tensor decomposition based unsupervised feature extraction Y-h. Taguchi, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | A lot of research has been done on the so-called textit{large p small n} problem, where the number of samples is small compared to the number of variables. In the so-called field of genome science, however, this ratio is extreme, with the number of genes (= number of variables = $p$) being tens of thousands while the number of subjects (= number of samples = $n$) is even a few, and $p/n sim 10^3$ is not uncommon. In such extreme cases, many of the methods proposed for the so-called textit{large p small n}problem are often ineffective. We have proposed ``Principal component analysis and tensor decomposition based unsupervised feature extraction'' to deal with this problem. In the past decade, we have applied this method to many researches in the field of bioinformatics. However, this method is purely within the scope of linear algebra, and since it is unsupervised learning, there is no tuning parameter, and if the method does not work, there is no choice but to give up on the analysis itself. In order to solve this problem, we have developed a kernel version of the method. In this paper, we report on how we succeeded in kernelizing the method to solve this problem, which enables the method to take nonlinear relationships into account and greatly expands its application range. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Tensor decomposition / Kernel trick / Feature selection / Unsupervised learning |
| Paper # | IBISML2020-36 |
| Date of Issue | 2021-02-23 (IBISML) |
| Conference Information | |
| Committee | IBISML |
|---|---|
| Conference Date | 2021/3/2(3days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Organized and general sessions on machine learning |
| Chair | Ichiro Takeuchi(Nagoya Inst. of Tech.) |
| Vice Chair | Masashi Sugiyama(Univ. of Tokyo) / Koji Tsuda(Univ. of Tokyo) |
| Secretary | Masashi Sugiyama(AIST) / Koji Tsuda(NTT) |
| Assistant | Atsuyoshi Nakamura(Hokkaido Univ.) / Shigeyuki Oba(Miidas) |
| Paper Information | |
| Registration To | Technical Committee on Infomation-Based Induction Sciences and Machine Learning |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Kernel tensor decomposition based unsupervised feature extraction |
| Sub Title (in English) | Applications to bioinformatics |
| Keyword(1) | Tensor decomposition |
| Keyword(2) | Kernel trick |
| Keyword(3) | Feature selection |
| Keyword(4) | Unsupervised learning |
| 1st Author's Name | Y-h. Taguchi |
| 1st Author's Affiliation | Chuo University(Chuo Univ.) |
| Date | 2021-03-02 |
| Paper # | IBISML2020-36 |
| Volume (vol) | vol.120 |
| Number (no) | IBISML-395 |
| Page | pp.pp.16-23(IBISML), |
| #Pages | 8 |
| Date of Issue | 2021-02-23 (IBISML) |