Social-Dimension Approach to Classification in Large-Scale Networks

Arizona State University, Computer Science and Engineering, Data Mining and Machine Learning

Overview

Social media often provides social network information of users. However, the relationship hidden in the connections are inhomogeneous. Social dimensions are introduced to represent the heterogeneous interactions people interact with each other (latent affiliations one actor is involved in). Here, we list some of our published work as well as the data sets and code used for experiments.

Data

Below are some of the data sets we have used for experiments in [2] and [3]. All are in matlab format. Each data set has two variables: network and groups. "network" is a symmetric sparse matrix representing the interaction between users, and "groups" are the groups subscribed by users. We use that as the class labels in our work.

BlogCatalog data of 10,312 nodes, 333, 983 links, and 39 categories.
Flickr Data of 80,513 nodes, 5, 899, 882 links, and 195 categories.
YouTube Data of 1, 138, 499 nodes, 2, 990, 443 links and 47 categories.

Code

SocioDim framework based on modularity maximization [2].
Extracting social dimensions based on EdgeClustering [3].

People

Tutorial

Huan Liu, Lei Tang, and Nitin Agarwal, "Tutorial: Community Detection and Behavior Study for Social Computing", presented at IEEE Infernational Conference on Social Computing (SocialCom'09), 2009.

References

[7] Lei Tang, Xufei Wang and Huan Liu. Scalable Learning of Collective Behavior. IEEE Transactions on Knowledge and Data Engineering (TKDE), To appear.

[6] Lei Tang, and Huan Liu. Leveraging Social Media Networks for Classification. Journal of Data Mining and Knowledge Discovery (DMKD), 2011.

[5] Lei Tang, Xufei Wang, Huan Liu and Lei Wang. A Multi-Resolution Approach to Learning with Overlapping Communities . In KDD Workshop on Social Media Analytics, 2010.

[4] Lei Tang and Huan Liu. Toward Predicting Collective Behavior via Social Dimension Extraction. In IEEE Intelligent Systems, 2010.

[3] Lei Tang and Huan Liu. Scalable Learning of Collective Behavior based on Sparse Social Dimensions. In Proceedings of the 18th ACM Conference on Information and Knowledge Management (CIKM’09), 2009.

[2] Lei Tang and Huan Liu. Relational Learning via Latent Social Dimensions. In Proceedings of The 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD’09), Pages 817–826, 2009.
[1] Lei Tang. Collective Behavior Prediction in Social Media. In Doctoral Student Forum, SIAM International Conference on Data Mining (SDM’09), 2009.

Other Related References

Aditya Krishna Menon and Charles Elkan. Predicting labels for dyadic data. In Proceedings of ECML, 2010.

Prithviraj Sen, Galileo Namata, Mustafa Bilgic, Lise Getoor, Biran Gallagher and Tina Eliassi-Rad. Collective Classification in Network Data. AI Magzine, 2008.

Sofus A. Macskassy and Foster Provost. Classification in Networked Data: A Toolkit and a Univariate Case Study, JMLR, 2007.

Acknowledgements

This project is sponsored by AFOSR-FA95500810132.