TY - GEN
T1 - Understanding animal behavior using their trajectories
T2 - 6th International Conference on Distributed, Ambient and Pervasive Interactions, DAPI 2018 Held as Part of HCI International 2018
AU - Ardakani, Ilya
AU - Hashimoto, Koichi
AU - Yoda, Ken
N1 - Funding Information:
Received June 21, 1984; revised Sept. 10, 1984; accepted Jan. 11, 1985. Drs. Satin and Sverd are with the Department of Psychiatry and Behavioral Science, State University of New York at Stony Brook. Drs. Winsberg and Foss are with the Nathan S. Kline Research Institute, New York State Office of Mental Health. Ms. Monetti is with the Bureau of Environmental Epidemiology and Occupational Health, New York State Department of Health. This work was supported by the New York State Office of Mental Health and a Basic Research Support Grant (907-EOlO) from the National Institutes of Health. The authors gratefully acknowledge the contributions of Nancy Young, R.N., who coordinated clinic procedures, Arthur Stone, Ph.D., who assisted in the diagnostic process and Benjamin Pasamnnick, M.D., and Jan hney, Ph.D., who provided insightful comments on an earlier draft. Reprints m y be requested from Dr. Bertrand Winsberg, Child Development Center, Nassau County Medical Center, 2201 Hemp-stead Turnpike, East Meadow, NY 11554 (516) 542-3948.
Publisher Copyright:
© Springer International Publishing AG, part of Springer Nature 2018.
PY - 2018
Y1 - 2018
N2 - Generally, behavior and movement could be closely attributed with each other. In other words, in most cases, behavior expressed as a set of specific movement patterns in time. These movement traces or trajectories representing behavior could provide a window into the underlying state of the subjects. In this study, analogies drawn between text and trajectories which allowed us to employ sentiment analysis and topic model methods to analyze trajectories. It is assumed that trajectories consist of key points which are commonly and frequently traversed. It is proposed that analogously in trajectory analysis, key points frequency would encapsulate information about the subject or the key points in trajectory generated by latent distribution which attributed to certain behavior or specific group of subjects with similar behavioral features. To test this hypothesis, an experiment was conducted which examines the influence of gender in composition of key points in birds’ trajectories logged from a seabird species called Streaked Shearwater Calonectris leucomelas. It was shown that genders have specific distribution over the key points. Therefore, key points membership in trajectory could be attributed to a specific gender and even a simple classifier would provide information about the gender of the subject simply by observing the trajectory’s key points. It was concluded that like text, trajectories composed of smaller elements which could be associated to a specific latent state. Learning or exploiting these associations revealed essential information about identity and behavior of the subject of observation.
AB - Generally, behavior and movement could be closely attributed with each other. In other words, in most cases, behavior expressed as a set of specific movement patterns in time. These movement traces or trajectories representing behavior could provide a window into the underlying state of the subjects. In this study, analogies drawn between text and trajectories which allowed us to employ sentiment analysis and topic model methods to analyze trajectories. It is assumed that trajectories consist of key points which are commonly and frequently traversed. It is proposed that analogously in trajectory analysis, key points frequency would encapsulate information about the subject or the key points in trajectory generated by latent distribution which attributed to certain behavior or specific group of subjects with similar behavioral features. To test this hypothesis, an experiment was conducted which examines the influence of gender in composition of key points in birds’ trajectories logged from a seabird species called Streaked Shearwater Calonectris leucomelas. It was shown that genders have specific distribution over the key points. Therefore, key points membership in trajectory could be attributed to a specific gender and even a simple classifier would provide information about the gender of the subject simply by observing the trajectory’s key points. It was concluded that like text, trajectories composed of smaller elements which could be associated to a specific latent state. Learning or exploiting these associations revealed essential information about identity and behavior of the subject of observation.
KW - Animal behavior
KW - Animal movement
KW - Trajectory mining
UR - http://www.scopus.com/inward/record.url?scp=85050555542&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85050555542&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-91131-1_1
DO - 10.1007/978-3-319-91131-1_1
M3 - Conference contribution
AN - SCOPUS:85050555542
SN - 9783319911304
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 3
EP - 22
BT - Distributed, Ambient and Pervasive Interactions
A2 - Konomi, Shin’ichi
A2 - Streitz, Norbert
PB - Springer Verlag
Y2 - 15 July 2018 through 20 July 2018
ER -