TY - GEN
T1 - Second-Order Asymptotically Optimal Change-point Detection Algorithm with Sampling Control
AU - Xu, Qunzhi
AU - Mei, Yajun
AU - Moustakides, George V.
N1 - Publisher Copyright:
© 2020 IEEE.
PY - 2020/6
Y1 - 2020/6
N2 - In the sequential change-point detection problem for multi-stream data, it is assumed that there are M processes in a system and at some unknown time, an occurring event impacts one unknown local process in the sense of changing the distribution of observations from that affected local process. In this paper, we consider such problem under the sampling control constraint, in which one is able to take observations from only one of the local processes at each time step. Our objective is to design an adaptive sampling policy and a stopping time policy that is able to raise a correct alarm as quickly as possible subject to the false alarm and sampling control constraint. We develop an efficient sequential change-point detection algorithm under the sampling control that turns out to be second-order asymptotically optimal under the full data scenario. That is, with the sampling rate that is only 1/M of the full data scenario, our proposed algorithm has the same performance up to second-order as the optimal procedure under the full data scenario.
AB - In the sequential change-point detection problem for multi-stream data, it is assumed that there are M processes in a system and at some unknown time, an occurring event impacts one unknown local process in the sense of changing the distribution of observations from that affected local process. In this paper, we consider such problem under the sampling control constraint, in which one is able to take observations from only one of the local processes at each time step. Our objective is to design an adaptive sampling policy and a stopping time policy that is able to raise a correct alarm as quickly as possible subject to the false alarm and sampling control constraint. We develop an efficient sequential change-point detection algorithm under the sampling control that turns out to be second-order asymptotically optimal under the full data scenario. That is, with the sampling rate that is only 1/M of the full data scenario, our proposed algorithm has the same performance up to second-order as the optimal procedure under the full data scenario.
KW - Asymptotic optimality
KW - change-point detection
KW - CUSUM
KW - quickest detection
UR - http://www.scopus.com/inward/record.url?scp=85090405589&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85090405589&partnerID=8YFLogxK
U2 - 10.1109/ISIT44484.2020.9174114
DO - 10.1109/ISIT44484.2020.9174114
M3 - Conference contribution
AN - SCOPUS:85090405589
T3 - IEEE International Symposium on Information Theory - Proceedings
SP - 1136
EP - 1140
BT - 2020 IEEE International Symposium on Information Theory, ISIT 2020 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2020 IEEE International Symposium on Information Theory, ISIT 2020
Y2 - 21 July 2020 through 26 July 2020
ER -