Social patterns

Community detection using behavior-generated network datasets

Alice Leung, William Dron, John P. Hancock, Matthew Aguirre, Jon Purnell, Jiawei Han, Chi Wang, Jaideep Srivastava, Amogh Mahapatra, Atanu Roy, Lisa Scott

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

1 Citation (Scopus)

Abstract

A set of behavior rules, personal characteristics, group affiliations and roles was used to generate a dataset of mixed communication actions modeling those at a large organization. Several different approaches to community detection and modeling were applied to this generated dataset, in order to compare the strengths and range of applicability of different algorithms. Graph partitioning methods performed well at assigning membership to formal, exclusive groups such as organizational departments, if there is a priori knowledge of the target number of groups. SSDE-cluster, a fast and scalable algorithm, performed well in detecting normal departments and can be used when the number of groups is not known. It also was able to detect small overlapping groups, but with only moderate accuracy. Clique enumeration performed well in detecting small overlapping groups, when a priori knowledge of average group size was used. Different methods of constructing social network graphs from the mixed communication actions were investigated, as well as different link weighing methods. We conclude that behavior-generated datasets with complex and complete ground truths are useful for collaborative validation of different community and role detection and modeling methods.

Original languageEnglish
Title of host publicationProceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013
Pages82-89
Number of pages8
DOIs
Publication statusPublished - 2013
Externally publishedYes
Event2013 IEEE 2nd International Network Science Workshop, NSW 2013 - West Point, NY
Duration: 29 Apr 20131 May 2013

Other

Other2013 IEEE 2nd International Network Science Workshop, NSW 2013
CityWest Point, NY
Period29/4/131/5/13

Fingerprint

Communication
Weighing

Keywords

  • clustering
  • communication models
  • community detection
  • generated datasets
  • group detection
  • rule-based behaviors

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Leung, A., Dron, W., Hancock, J. P., Aguirre, M., Purnell, J., Han, J., ... Scott, L. (2013). Social patterns: Community detection using behavior-generated network datasets. In Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013 (pp. 82-89). [6609198] https://doi.org/10.1109/NSW.2013.6609198

Social patterns : Community detection using behavior-generated network datasets. / Leung, Alice; Dron, William; Hancock, John P.; Aguirre, Matthew; Purnell, Jon; Han, Jiawei; Wang, Chi; Srivastava, Jaideep; Mahapatra, Amogh; Roy, Atanu; Scott, Lisa.

Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013. 2013. p. 82-89 6609198.

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

Leung, A, Dron, W, Hancock, JP, Aguirre, M, Purnell, J, Han, J, Wang, C, Srivastava, J, Mahapatra, A, Roy, A & Scott, L 2013, Social patterns: Community detection using behavior-generated network datasets. in Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013., 6609198, pp. 82-89, 2013 IEEE 2nd International Network Science Workshop, NSW 2013, West Point, NY, 29/4/13. https://doi.org/10.1109/NSW.2013.6609198
Leung A, Dron W, Hancock JP, Aguirre M, Purnell J, Han J et al. Social patterns: Community detection using behavior-generated network datasets. In Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013. 2013. p. 82-89. 6609198 https://doi.org/10.1109/NSW.2013.6609198
Leung, Alice ; Dron, William ; Hancock, John P. ; Aguirre, Matthew ; Purnell, Jon ; Han, Jiawei ; Wang, Chi ; Srivastava, Jaideep ; Mahapatra, Amogh ; Roy, Atanu ; Scott, Lisa. / Social patterns : Community detection using behavior-generated network datasets. Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013. 2013. pp. 82-89
@inproceedings{a58ba1169e614aec990e6baca7db645c,
title = "Social patterns: Community detection using behavior-generated network datasets",
abstract = "A set of behavior rules, personal characteristics, group affiliations and roles was used to generate a dataset of mixed communication actions modeling those at a large organization. Several different approaches to community detection and modeling were applied to this generated dataset, in order to compare the strengths and range of applicability of different algorithms. Graph partitioning methods performed well at assigning membership to formal, exclusive groups such as organizational departments, if there is a priori knowledge of the target number of groups. SSDE-cluster, a fast and scalable algorithm, performed well in detecting normal departments and can be used when the number of groups is not known. It also was able to detect small overlapping groups, but with only moderate accuracy. Clique enumeration performed well in detecting small overlapping groups, when a priori knowledge of average group size was used. Different methods of constructing social network graphs from the mixed communication actions were investigated, as well as different link weighing methods. We conclude that behavior-generated datasets with complex and complete ground truths are useful for collaborative validation of different community and role detection and modeling methods.",
keywords = "clustering, communication models, community detection, generated datasets, group detection, rule-based behaviors",
author = "Alice Leung and William Dron and Hancock, {John P.} and Matthew Aguirre and Jon Purnell and Jiawei Han and Chi Wang and Jaideep Srivastava and Amogh Mahapatra and Atanu Roy and Lisa Scott",
year = "2013",
doi = "10.1109/NSW.2013.6609198",
language = "English",
isbn = "9781479904365",
pages = "82--89",
booktitle = "Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013",

}

TY - GEN

T1 - Social patterns

T2 - Community detection using behavior-generated network datasets

AU - Leung, Alice

AU - Dron, William

AU - Hancock, John P.

AU - Aguirre, Matthew

AU - Purnell, Jon

AU - Han, Jiawei

AU - Wang, Chi

AU - Srivastava, Jaideep

AU - Mahapatra, Amogh

AU - Roy, Atanu

AU - Scott, Lisa

PY - 2013

Y1 - 2013

N2 - A set of behavior rules, personal characteristics, group affiliations and roles was used to generate a dataset of mixed communication actions modeling those at a large organization. Several different approaches to community detection and modeling were applied to this generated dataset, in order to compare the strengths and range of applicability of different algorithms. Graph partitioning methods performed well at assigning membership to formal, exclusive groups such as organizational departments, if there is a priori knowledge of the target number of groups. SSDE-cluster, a fast and scalable algorithm, performed well in detecting normal departments and can be used when the number of groups is not known. It also was able to detect small overlapping groups, but with only moderate accuracy. Clique enumeration performed well in detecting small overlapping groups, when a priori knowledge of average group size was used. Different methods of constructing social network graphs from the mixed communication actions were investigated, as well as different link weighing methods. We conclude that behavior-generated datasets with complex and complete ground truths are useful for collaborative validation of different community and role detection and modeling methods.

AB - A set of behavior rules, personal characteristics, group affiliations and roles was used to generate a dataset of mixed communication actions modeling those at a large organization. Several different approaches to community detection and modeling were applied to this generated dataset, in order to compare the strengths and range of applicability of different algorithms. Graph partitioning methods performed well at assigning membership to formal, exclusive groups such as organizational departments, if there is a priori knowledge of the target number of groups. SSDE-cluster, a fast and scalable algorithm, performed well in detecting normal departments and can be used when the number of groups is not known. It also was able to detect small overlapping groups, but with only moderate accuracy. Clique enumeration performed well in detecting small overlapping groups, when a priori knowledge of average group size was used. Different methods of constructing social network graphs from the mixed communication actions were investigated, as well as different link weighing methods. We conclude that behavior-generated datasets with complex and complete ground truths are useful for collaborative validation of different community and role detection and modeling methods.

KW - clustering

KW - communication models

KW - community detection

KW - generated datasets

KW - group detection

KW - rule-based behaviors

UR - http://www.scopus.com/inward/record.url?scp=84886067103&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84886067103&partnerID=8YFLogxK

U2 - 10.1109/NSW.2013.6609198

DO - 10.1109/NSW.2013.6609198

M3 - Conference contribution

SN - 9781479904365

SP - 82

EP - 89

BT - Proceedings of the 2013 IEEE 2nd International Network Science Workshop, NSW 2013

ER -