| 講演抄録/キーワード |
| 講演名 |
2011-11-10 15:45
Change-Point Detection in Time-Series Data by Relative Density-Ratio Estimation ○Song Liu・Makoto Yamada・Masashi Sugiyama(Tokyo Inst. of Tech.) IBISML2011-70 |
| 抄録 |
(和) |
The objective of change-point detection is to discover abrupt property c
hanges lying behind time series data. In this paper, we present a novel
statistical change-point detection algorithm that is based on non-parame
tric divergence estimation between two retrospective segments. Our metho
d uses the relative Pearson divergence as a divergence measure, and it i
s accurately and efficiently estimated by a method of direct density-rat
io estimation. Through experiments on artificial and real-world datasets
including human-activity sensing, speeches, and Twitter archives, we de
monstrate the usefulness of the proposed method. |
| (英) |
The objective of change-point detection is to discover abrupt property changes lying behind time series data. In this paper, we present a novel statistical change-point detection algorithm that is based on non-parametric divergence estimation between two retrospective segments. Our method uses the relative Pearson divergence as a divergence measure, and it is accurately and efficiently estimated by a method of direct density-ratio estimation. Through experiments on artificial and real-world datasets including human-activity sensing, speeches, and Twitter archives, we demonstrate the usefulness of the proposed method. |
| キーワード |
(和) |
/ / / / / / / |
| (英) |
change-point detection / distribution comparison / relative density-ratio estimation / kernel methods / time-series data / / / |
| 文献情報 |
信学技報, vol. 111, no. 275, IBISML2011-70, pp. 187-198, 2011年11月. |
| 資料番号 |
IBISML2011-70 |
| 発行日 |
2011-11-02 (IBISML) |
| ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
| PDFダウンロード |
IBISML2011-70 |