Transcript Dynamics of communities in two fission

Dynamics of communities in two fission-fusion species, Grevy's zebra and onager
Chayant Tantipathananandh1, Tanya Y. Berger-Wolf1, Siva R. Sundaresan2, Ilya R. Fischhoff2, and Daniel I. Rubenstein2
1Dept. of Computer Science, University of Illinois at Chicago, USA
2Dept. of Ecology and Evolutionary Biology, Princeton University, USA
Grevy’s zebras
Introduction
Onagers
Static Networks
We consider the problem of identifying communities in populations with changing social affiliations over time. Ours is the
first community identification method that does not aggregate information over time but takes the dynamic nature of
interactions explicitly into consideration, accounting for a fluid and changing community structure.
Here we apply our method to two equid species: Grevy's zebra (Equus grevyi) and onagers (Equus hemionus khur). The two
populations are of similar fission-fusion species with territorial males and harem females. However, there are ecological
differences which have been expected to result in different social patterns between the two species. The results of applying
our method not only confirm the ecological insight but quantify it and provide visual supporting evidence of the difference. In
particular, Grevy's zebra have a more cohesive social structure, which suggests a preference of individual company. In
contrast, the social groups in onagers are more happenstance, which suggests that group formation is a byproduct of other
ecological factors, such as congregation at a waterhole or seeking strength in numbers against predation
The recurring
community
(see below)
Observation
We assume that the input is obtained as a sequence of population snapshots.
Day 1
1
2
Groups of individuals in each snapshot
are defined by spatial proximity
3
5
T1
4
Day 3
Dynamic Networks
2
5
T2
5
T3
5
T4
5
T5
5
4
2
2
2 3 1
3
3
4
1
4
1
Unseen
Individuals
A small
community
4
1
2
2
3
4
1
4
3 1
time
5
Observed
Individuals
A recurring
community
Fleeting small
communities,
otherwise isolated
individuals.
Dynamic Community Identification Method
A persistent
community
We assume that social individuals need to maintain the following properties:
• Individual Conservatism They rarely change community affiliations
• Group Loyalty
• They tend to stay near by other individuals from the same community to socialize
• They tend to be unseen at the same time
Deviations from these assumptions are possible and ideally need to be at minimum to maintain socialization. Finding
dynamic
community
that
minimizes
these
deviations
is
NP-complete.
Example Analysis
Affiliation Change:
Green-blue-green
fission
Green individual in
blue group
Green individual
absent from green group
Entire group
was unseen
fusion
fission
Figure 2. Each color represents one community. Color of a group indicates which community the group represents
(which may or may not be dominating color within group). Color of individuals indicate its affiliation at the moment.
fusion
References
Sundaresan, S. R., Fischhoff, I. R., Dushoff, J. & Rubenstein, D. I. 2007. Network metrics reveal differences in social
organization between two fission-fusion species, Grevy's zebra and onager. Oecologia, 151, 140-149.
Tantipathananandh, C., Berger-Wolf T., Kempe, D. 2007. A Framework For Identifying Communities in Dynamic Social
Networks. Proceedings of the 13th ACM SIGKDD, 717-726.
fission