TY - GEN
T1 - Information-space partitioning and symbolization of multi-dimensional time-series data using density estimation
AU - Virani, Nurali
AU - Lee, Ji Woong
AU - Phoha, Shashi
AU - Ray, Asok
N1 - Funding Information:
This work was supported in part by the U.S. Air Force Office of Scientific Research (AFOSR) under Grant No. FA9550-12-1-0270, and by the National Science Foundation (NSF) under Grant No. ECCS-1201973. Any opinions, findings, and conclusions or recommendations expressed in this paper are those of the authors and do not necessarily reflect the views of the sponsoring agencies
Publisher Copyright:
© 2016 American Automatic Control Council (AACC).
PY - 2016/7/28
Y1 - 2016/7/28
N2 - This paper proposes a nonparametric density estimation-based information-space partitioning and symbolization technique for capturing and representing the underlying statistical behavior in dynamic data-driven application systems (DDDAS). In contrast with existing tools that address alphabet-size selection and partitioning in two separate steps, the proposed technique jointly determines both the number of symbols (i.e., the alphabet size) and the regions associated with these symbols (i.e., the information-space partition) in a single step. In order to validate the technique, dynamic data-driven models have been developed from time series of vector-valued measurements extracted from a simulation testbed, and are compared with models that rely on Gaussian mixture modeling for information-space partitioning.
AB - This paper proposes a nonparametric density estimation-based information-space partitioning and symbolization technique for capturing and representing the underlying statistical behavior in dynamic data-driven application systems (DDDAS). In contrast with existing tools that address alphabet-size selection and partitioning in two separate steps, the proposed technique jointly determines both the number of symbols (i.e., the alphabet size) and the regions associated with these symbols (i.e., the information-space partition) in a single step. In order to validate the technique, dynamic data-driven models have been developed from time series of vector-valued measurements extracted from a simulation testbed, and are compared with models that rely on Gaussian mixture modeling for information-space partitioning.
UR - http://www.scopus.com/inward/record.url?scp=84992035051&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84992035051&partnerID=8YFLogxK
U2 - 10.1109/ACC.2016.7525431
DO - 10.1109/ACC.2016.7525431
M3 - Conference contribution
AN - SCOPUS:84992035051
T3 - Proceedings of the American Control Conference
SP - 3328
EP - 3333
BT - 2016 American Control Conference, ACC 2016
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 2016 American Control Conference, ACC 2016
Y2 - 6 July 2016 through 8 July 2016
ER -