| Presentation | 2020-05-28 Bayes Optimal Detecting Relevant Changes for i.p.i.d. Sources Kairi Suzuki, Akira Kamatsuka, Toshiyasu Matsushima, |
|---|---|
| PDF Download Page |  PDF download Page Link |
| Abstract(in Japanese) | (See Japanese page) |
| Abstract(in English) | The problems of detecting change points are studied in various fields. There are various types of change-point detection-related problems such as detecting whether the change has occured, estimating on magnitude of the change $abs{theta_{t_{0}}-theta_{t}}$, and estimating on the number of the changes. In most of the previous studies, problems of detecting the change point are formulated via hypothesis testing of the form $H_{0}:theta_{t}=theta_{t_{0}}$ v.s. $H_{1}: theta_{t} neq theta_{t_{0}}$. Recently Dette et al. has formualted the problemvia hypothesis testing of the form $abs{theta_{t}-theta_{t_{0}}}< r$ v.s. $H_{1}:abs{theta_{t}-theta_{t_{0}}} geq r$, which ignores small changes, where $r$ is a constant. The change point problem of this form is called a relevent-change-point detection problem.. In this study, we formulate a detecting the relevant change-points in the statistical decision theory. Then we propose an optimal and efficient estimating algorithm of the number of change points under bayesian criterion. The algorithm take the computational complexity in the order of $O(n^{3})$ under some assumptions. |
| Keyword(in Japanese) | (See Japanese page) |
| Keyword(in English) | Change point detection / Relevant changes / i.p.i.d. sources / Bayesian decision theory |
| Paper # | IT2020-3,EMM2020-3 |
| Date of Issue | 2020-05-21 (IT, EMM) |
| Conference Information | |
| Committee | IT / EMM |
|---|---|
| Conference Date | 2020/5/28(2days) |
| Place (in Japanese) | (See Japanese page) |
| Place (in English) | Online |
| Topics (in Japanese) | (See Japanese page) |
| Topics (in English) | Information Security, Information Theory, Information Hiding, etc. |
| Chair | Jun Muramatsu(NTT) / Masaki Kawamura(Yamaguchi Univ.) |
| Vice Chair | Tadashi Wadayama(Nagoya Inst. of Tech.) / Motoi Iwata(Osaka Prefecture Univ.) / Tetsuya Kojima(NIT,Tokyo College) |
| Secretary | Tadashi Wadayama(Saga Univ.) / Motoi Iwata(Senshu University) / Tetsuya Kojima(NIT, Nagano College) |
| Assistant | Hideki Yagi(UEC) / Masaki Inamura(Tokyo Denki Univ.) / Kazuhiro Kono(Kansai Univ.) |
| Paper Information | |
| Registration To | Technical Committee on Information Theory / Technical Committee on Enriched MultiMedia |
|---|---|
| Language | JPN |
| Title (in Japanese) | (See Japanese page) |
| Sub Title (in Japanese) | (See Japanese page) |
| Title (in English) | Bayes Optimal Detecting Relevant Changes for i.p.i.d. Sources |
| Sub Title (in English) | |
| Keyword(1) | Change point detection |
| Keyword(2) | Relevant changes |
| Keyword(3) | i.p.i.d. sources |
| Keyword(4) | Bayesian decision theory |
| 1st Author's Name | Kairi Suzuki |
| 1st Author's Affiliation | Waseda University(Waseda Univ.) |
| 2nd Author's Name | Akira Kamatsuka |
| 2nd Author's Affiliation | Waseda University(Waseda Univ.) |
| 3rd Author's Name | Toshiyasu Matsushima |
| 3rd Author's Affiliation | Waseda University(Waseda Univ.) |
| Date | 2020-05-28 |
| Paper # | IT2020-3,EMM2020-3 |
| Volume (vol) | vol.120 |
| Number (no) | IT-43,EMM-42 |
| Page | pp.pp.13-18(IT), pp.13-18(EMM), |
| #Pages | 6 |
| Date of Issue | 2020-05-21 (IT, EMM) |