Role-behavior analysis from trajectory data by cross-domain learning

Shin Ando, Einoshin Suzuki

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Behavior analysis using trajectory data presents a practical and interesting challenge for KDD. Conventional analyses address discriminative tasks of behaviors, e.g., classification and clustering typically using the subsequences extracted from the trajectory of an object as a numerical feature representation. In this paper, we explore further to identify the difference in the high-level semantics of behaviors such as roles and address the task in a cross-domain learning approach. The trajectory, from which the features are sampled, is intuitively viewed as a domain, and we assume that its intrinsic structure is characterized by the underlying role associated with the tracked object. We propose a novel hybrid method of spectral clustering and density approximation for comparing clustering structures of two independently sampled trajectory data and identifying patterns of behaviors unique to a role. We present empirical evaluations of the proposed method in two practical settings using real-world robotic trajectories.

Original languageEnglish
Title of host publicationProceedings - 11th IEEE International Conference on Data Mining, ICDM 2011
Pages21-30
Number of pages10
DOIs
Publication statusPublished - Dec 1 2011
Event11th IEEE International Conference on Data Mining, ICDM 2011 - Vancouver, BC, Canada
Duration: Dec 11 2011Dec 14 2011

Other

Other11th IEEE International Conference on Data Mining, ICDM 2011
CountryCanada
CityVancouver, BC
Period12/11/1112/14/11

Fingerprint

Trajectories
Robotics
Semantics

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Cite this

Ando, S., & Suzuki, E. (2011). Role-behavior analysis from trajectory data by cross-domain learning. In Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011 (pp. 21-30). [6137206] https://doi.org/10.1109/ICDM.2011.125

Role-behavior analysis from trajectory data by cross-domain learning. / Ando, Shin; Suzuki, Einoshin.

Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011. 2011. p. 21-30 6137206.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Ando, S & Suzuki, E 2011, Role-behavior analysis from trajectory data by cross-domain learning. in Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011., 6137206, pp. 21-30, 11th IEEE International Conference on Data Mining, ICDM 2011, Vancouver, BC, Canada, 12/11/11. https://doi.org/10.1109/ICDM.2011.125
Ando S, Suzuki E. Role-behavior analysis from trajectory data by cross-domain learning. In Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011. 2011. p. 21-30. 6137206 https://doi.org/10.1109/ICDM.2011.125
Ando, Shin ; Suzuki, Einoshin. / Role-behavior analysis from trajectory data by cross-domain learning. Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011. 2011. pp. 21-30
@inproceedings{604098603b7244f39e75d03e569d7476,
title = "Role-behavior analysis from trajectory data by cross-domain learning",
abstract = "Behavior analysis using trajectory data presents a practical and interesting challenge for KDD. Conventional analyses address discriminative tasks of behaviors, e.g., classification and clustering typically using the subsequences extracted from the trajectory of an object as a numerical feature representation. In this paper, we explore further to identify the difference in the high-level semantics of behaviors such as roles and address the task in a cross-domain learning approach. The trajectory, from which the features are sampled, is intuitively viewed as a domain, and we assume that its intrinsic structure is characterized by the underlying role associated with the tracked object. We propose a novel hybrid method of spectral clustering and density approximation for comparing clustering structures of two independently sampled trajectory data and identifying patterns of behaviors unique to a role. We present empirical evaluations of the proposed method in two practical settings using real-world robotic trajectories.",
author = "Shin Ando and Einoshin Suzuki",
year = "2011",
month = "12",
day = "1",
doi = "10.1109/ICDM.2011.125",
language = "English",
isbn = "9780769544083",
pages = "21--30",
booktitle = "Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011",

}

TY - GEN

T1 - Role-behavior analysis from trajectory data by cross-domain learning

AU - Ando, Shin

AU - Suzuki, Einoshin

PY - 2011/12/1

Y1 - 2011/12/1

N2 - Behavior analysis using trajectory data presents a practical and interesting challenge for KDD. Conventional analyses address discriminative tasks of behaviors, e.g., classification and clustering typically using the subsequences extracted from the trajectory of an object as a numerical feature representation. In this paper, we explore further to identify the difference in the high-level semantics of behaviors such as roles and address the task in a cross-domain learning approach. The trajectory, from which the features are sampled, is intuitively viewed as a domain, and we assume that its intrinsic structure is characterized by the underlying role associated with the tracked object. We propose a novel hybrid method of spectral clustering and density approximation for comparing clustering structures of two independently sampled trajectory data and identifying patterns of behaviors unique to a role. We present empirical evaluations of the proposed method in two practical settings using real-world robotic trajectories.

AB - Behavior analysis using trajectory data presents a practical and interesting challenge for KDD. Conventional analyses address discriminative tasks of behaviors, e.g., classification and clustering typically using the subsequences extracted from the trajectory of an object as a numerical feature representation. In this paper, we explore further to identify the difference in the high-level semantics of behaviors such as roles and address the task in a cross-domain learning approach. The trajectory, from which the features are sampled, is intuitively viewed as a domain, and we assume that its intrinsic structure is characterized by the underlying role associated with the tracked object. We propose a novel hybrid method of spectral clustering and density approximation for comparing clustering structures of two independently sampled trajectory data and identifying patterns of behaviors unique to a role. We present empirical evaluations of the proposed method in two practical settings using real-world robotic trajectories.

UR - http://www.scopus.com/inward/record.url?scp=84857159764&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857159764&partnerID=8YFLogxK

U2 - 10.1109/ICDM.2011.125

DO - 10.1109/ICDM.2011.125

M3 - Conference contribution

AN - SCOPUS:84857159764

SN - 9780769544083

SP - 21

EP - 30

BT - Proceedings - 11th IEEE International Conference on Data Mining, ICDM 2011

ER -