TY - GEN
T1 - Facets
T2 - 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD 2015
AU - Cai, Yongjie
AU - Tong, Hanghang
AU - Fan, Wei
AU - Ji, Ping
AU - He, Qing
N1 - Publisher Copyright:
© 2015 ACM.
PY - 2015/8/10
Y1 - 2015/8/10
N2 - Mining time series data has been a very active research area in the past decade, exactly because of its prevalence in many high-impact applications, ranging from environmental monitoring, intelligent transportation systems, computer network forensics, to smart buildings and many more. It has posed many fascinating research questions. Among others, three prominent challenges shared by a variety of real applications are (a) high-order; (b) contextual constraints and (c) temporal smoothness. The state-of-the-art mining algorithms are rich in addressing each of these challenges, but relatively short of comprehensiveness in attacking the coexistence of multiple or even all of these three challenges. In this paper, we propose a comprehensive method, FACETS, to simultaneously model all these three challenges. We formulate it as an optimization problem from a dynamic graphical model perspective. The key idea is to use tensor factorization to address multi-aspect challenges, and perform careful regularizations to attack both contextual and temporal challenges. Based on that, we propose an effective and scalable algorithm to solve the problem. Our experimental evaluations on three real datasets demonstrate that our method (1) outperforms its competitors in two common data mining tasks (imputation and prediction); and (2) enjoys a linear scalability w.r.t. the length of time series.
AB - Mining time series data has been a very active research area in the past decade, exactly because of its prevalence in many high-impact applications, ranging from environmental monitoring, intelligent transportation systems, computer network forensics, to smart buildings and many more. It has posed many fascinating research questions. Among others, three prominent challenges shared by a variety of real applications are (a) high-order; (b) contextual constraints and (c) temporal smoothness. The state-of-the-art mining algorithms are rich in addressing each of these challenges, but relatively short of comprehensiveness in attacking the coexistence of multiple or even all of these three challenges. In this paper, we propose a comprehensive method, FACETS, to simultaneously model all these three challenges. We formulate it as an optimization problem from a dynamic graphical model perspective. The key idea is to use tensor factorization to address multi-aspect challenges, and perform careful regularizations to attack both contextual and temporal challenges. Based on that, we propose an effective and scalable algorithm to solve the problem. Our experimental evaluations on three real datasets demonstrate that our method (1) outperforms its competitors in two common data mining tasks (imputation and prediction); and (2) enjoys a linear scalability w.r.t. the length of time series.
KW - A network of time series
KW - Tensor factorization
UR - http://www.scopus.com/inward/record.url?scp=84954190187&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84954190187&partnerID=8YFLogxK
U2 - 10.1145/2783258.2783348
DO - 10.1145/2783258.2783348
M3 - Conference contribution
AN - SCOPUS:84954190187
T3 - Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining
SP - 79
EP - 88
BT - KDD 2015 - Proceedings of the 21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining
PB - Association for Computing Machinery
Y2 - 10 August 2015 through 13 August 2015
ER -