Paper Abstract and Keywords |
Presentation |
2016-09-05 13:15
[Short Paper]
Selective Inference for Time-series Change-Point Analysis Yuta Umezu, Kazuya Nakagawa, Shigenori Inoue (NIT), Koji Tsuda (Tokyo Univ.), Mahito Sugiyama, Takuya Maekawa (Osaka Univ.), Toru Tamaki (Hiroshima Univ.), Ken Yoda (Nagoya Univ.), Ichiro Takeuchi (NIT) PRMU2016-63 IBISML2016-18 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we propose a statistical method for time series data after detecting a change point. Because the change point is detected based on the data, its significance is highly depend on an algorithm for change point detection. Therefore, we need considering the algorithm to conduct a statistical inference. This kind of statistical inference is also known as post selection inference. We use the so-called cumulative sum statistic for detecting a change point, and then we conduct a Selective Inference based on the statistic. Through simulation studies, we confirm that the our proposal is almost better than existing methods. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
change point detection / CUSUM test / hypothesis testing / selective inference / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 116, no. 209, IBISML2016-18, pp. 89-92, Sept. 2016. |
Paper # |
IBISML2016-18 |
Date of Issue |
2016-08-29 (PRMU, IBISML) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
PRMU2016-63 IBISML2016-18 |
|