WICER

A weighted inter-cluster edge ranking for clustered graphs

Divya Padmanabhan, Prasanna Desikan, Jaideep Srivastava, Kashif Riaz

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

10 Citations (Scopus)

Abstract

Several algorithms based on link analysis have been developed to measure the importance of nodes on a graph such as pages on the World Wide Web. PageRank and HITS are the most popular ranking algorithms to rank the nodes of any directed graph. But, both these algorithms assign equal importance to all the edges and nodes, ignoring the semantically rich information from nodes and edges. Therefore, in the case of a graph containing natural clusters, these algorithms do not differentiate between inter-cluster edges and infra-cluster edges. Based on this parameter, we propose a Weighted InterCluster Edge Ranking for clustered graphs that weighs edges (based on whether it is an inter-cluster or an infracluster edge) and nodes (based on the number of clusters it connects). We introduce a parameter 'a' which can be adjusted depending on the bias desired in a clustered graph. Our experiments were two fold. We implemented our algorithm to relationship set representing legal entities and documents and the results indicate the significance of the weighted edge approach. We also generated biased and random walks to quantitatively study the performance.

Original languageEnglish
Title of host publicationProceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005
Pages522-528
Number of pages7
Volume2005
DOIs
Publication statusPublished - 2005
Externally publishedYes
Event2005 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2005 - Compiegne Cedex,
Duration: 19 Sep 200522 Sep 2005

Other

Other2005 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2005
CityCompiegne Cedex,
Period19/9/0522/9/05

Fingerprint

Directed graphs
Experiments

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Padmanabhan, D., Desikan, P., Srivastava, J., & Riaz, K. (2005). WICER: A weighted inter-cluster edge ranking for clustered graphs. In Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005 (Vol. 2005, pp. 522-528). [1517903] https://doi.org/10.1109/WI.2005.166

WICER : A weighted inter-cluster edge ranking for clustered graphs. / Padmanabhan, Divya; Desikan, Prasanna; Srivastava, Jaideep; Riaz, Kashif.

Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005. Vol. 2005 2005. p. 522-528 1517903.

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

Padmanabhan, D, Desikan, P, Srivastava, J & Riaz, K 2005, WICER: A weighted inter-cluster edge ranking for clustered graphs. in Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005. vol. 2005, 1517903, pp. 522-528, 2005 IEEE/WIC/ACM International Conference on Web Intelligence, WI 2005, Compiegne Cedex, 19/9/05. https://doi.org/10.1109/WI.2005.166
Padmanabhan D, Desikan P, Srivastava J, Riaz K. WICER: A weighted inter-cluster edge ranking for clustered graphs. In Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005. Vol. 2005. 2005. p. 522-528. 1517903 https://doi.org/10.1109/WI.2005.166
Padmanabhan, Divya ; Desikan, Prasanna ; Srivastava, Jaideep ; Riaz, Kashif. / WICER : A weighted inter-cluster edge ranking for clustered graphs. Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005. Vol. 2005 2005. pp. 522-528
@inproceedings{24f8246043e749dca77744ffacaa43eb,
title = "WICER: A weighted inter-cluster edge ranking for clustered graphs",
abstract = "Several algorithms based on link analysis have been developed to measure the importance of nodes on a graph such as pages on the World Wide Web. PageRank and HITS are the most popular ranking algorithms to rank the nodes of any directed graph. But, both these algorithms assign equal importance to all the edges and nodes, ignoring the semantically rich information from nodes and edges. Therefore, in the case of a graph containing natural clusters, these algorithms do not differentiate between inter-cluster edges and infra-cluster edges. Based on this parameter, we propose a Weighted InterCluster Edge Ranking for clustered graphs that weighs edges (based on whether it is an inter-cluster or an infracluster edge) and nodes (based on the number of clusters it connects). We introduce a parameter 'a' which can be adjusted depending on the bias desired in a clustered graph. Our experiments were two fold. We implemented our algorithm to relationship set representing legal entities and documents and the results indicate the significance of the weighted edge approach. We also generated biased and random walks to quantitatively study the performance.",
author = "Divya Padmanabhan and Prasanna Desikan and Jaideep Srivastava and Kashif Riaz",
year = "2005",
doi = "10.1109/WI.2005.166",
language = "English",
isbn = "076952415X",
volume = "2005",
pages = "522--528",
booktitle = "Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005",

}

TY - GEN

T1 - WICER

T2 - A weighted inter-cluster edge ranking for clustered graphs

AU - Padmanabhan, Divya

AU - Desikan, Prasanna

AU - Srivastava, Jaideep

AU - Riaz, Kashif

PY - 2005

Y1 - 2005

N2 - Several algorithms based on link analysis have been developed to measure the importance of nodes on a graph such as pages on the World Wide Web. PageRank and HITS are the most popular ranking algorithms to rank the nodes of any directed graph. But, both these algorithms assign equal importance to all the edges and nodes, ignoring the semantically rich information from nodes and edges. Therefore, in the case of a graph containing natural clusters, these algorithms do not differentiate between inter-cluster edges and infra-cluster edges. Based on this parameter, we propose a Weighted InterCluster Edge Ranking for clustered graphs that weighs edges (based on whether it is an inter-cluster or an infracluster edge) and nodes (based on the number of clusters it connects). We introduce a parameter 'a' which can be adjusted depending on the bias desired in a clustered graph. Our experiments were two fold. We implemented our algorithm to relationship set representing legal entities and documents and the results indicate the significance of the weighted edge approach. We also generated biased and random walks to quantitatively study the performance.

AB - Several algorithms based on link analysis have been developed to measure the importance of nodes on a graph such as pages on the World Wide Web. PageRank and HITS are the most popular ranking algorithms to rank the nodes of any directed graph. But, both these algorithms assign equal importance to all the edges and nodes, ignoring the semantically rich information from nodes and edges. Therefore, in the case of a graph containing natural clusters, these algorithms do not differentiate between inter-cluster edges and infra-cluster edges. Based on this parameter, we propose a Weighted InterCluster Edge Ranking for clustered graphs that weighs edges (based on whether it is an inter-cluster or an infracluster edge) and nodes (based on the number of clusters it connects). We introduce a parameter 'a' which can be adjusted depending on the bias desired in a clustered graph. Our experiments were two fold. We implemented our algorithm to relationship set representing legal entities and documents and the results indicate the significance of the weighted edge approach. We also generated biased and random walks to quantitatively study the performance.

UR - http://www.scopus.com/inward/record.url?scp=33748879611&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33748879611&partnerID=8YFLogxK

U2 - 10.1109/WI.2005.166

DO - 10.1109/WI.2005.166

M3 - Conference contribution

SN - 076952415X

SN - 9780769524153

VL - 2005

SP - 522

EP - 528

BT - Proceedings - 2005 IEEE/WIC/ACM InternationalConference on Web Intelligence, WI 2005

ER -