AI helps you reading Science

AI generates interpretation videos

AI extracts and analyses the key points of the paper to generate videos automatically


pub
Go Generating

AI Traceability

AI parses the academic lineage of this thesis


Master Reading Tree
Generate MRT

AI Insight

AI extracts a summary of this paper


Weibo:
Experimental results show that by simultaneously considering social network structure as well as user profile information our method performs significantly better than natural alternatives and the current state-of-the-art

Learning to Discover Social Circles in Ego Networks.

NIPS, pp.548-556, (2012)

Cited by: 1781|Views130
EI
Full Text
Bibtex
Weibo

Abstract

Our personal social networks are big and cluttered, and currently there is no good way to organize them. Social networking sites allow users to manually categorize their friends into social circles (e.g. ‘circles’ on Google+, and ‘lists’ on Facebook and Twitter), however they are laborious to construct and must be updated whenever a user’...More

Code:

Data:

0
Introduction
  • Online social networks allow users to follow streams of posts generated by hundreds of their friends and acquaintances.
  • The authors learn the model by simultaneously choosing node circle memberships and profile similarity functions so as to best explain the observed data.
  • The authors introduce a dataset of 1,143 ego-networks from Facebook, Google+, and Twitter, for which the authors obtain hand-labeled ground-truth from 5,636 different circles.1 Experimental results show that by simultaneously considering social network structure as well as user profile information the method performs significantly better than natural alternatives and the current state-of-the-art.
Highlights
  • Online social networks allow users to follow streams of posts generated by hundreds of their friends and acquaintances
  • Given a single user with her personal social network, our goal is to identify her circles, each of which is a subset of her friends
  • Experimental results show that by simultaneously considering social network structure as well as user profile information our method performs significantly better than natural alternatives and the current state-of-the-art
  • We describe a model of social circles that treats circle memberships as latent variables
  • We considered a wide number of baseline methods, including those that consider only network structure, those that consider only profile information, and those that consider both
Results
  • The authors' work is related to [30] in the sense that it performs clustering on social-network data, and [23], which models memberships to multiple communities.
  • The authors describe how to optimize node circle memberships C as well as the parameters of the user profile similarity functions Θ = {} (k = 1 .
  • This means that the method could be run on the full Facebook graph, as circles are independently detected for each user, and the ego-networks typically contain only hundreds of nodes.
  • From Facebook the authors obtained profile and network data from 10 ego-networks, consisting of 193 circles and 4,039 users.
  • From Google+ the authors obtained data from 133 ego-networks, consisting of 479 circles and 106,674 users.
  • From Twitter the authors obtained data from 1,000 ego-networks, consisting of 4,869 circles and 81,362 users.
  • Taken together the data contains 1,143 different ego-networks, 5,541 circles, and 192,075 users.
  • Note that feature descriptors are defined per ego-network: while many thousands of high schools exist among all Facebook users, only a small number appear among any particular user’s friends.
  • One way to address this is to form difference vectors based on the parents of leaf nodes: this way, the authors encode what profile categories two users have in common, but disregard specific values (Figure 2, bottom right).
Conclusion
  • Mixedmembership models predict a stochastic vector encoding partial circle memberships, which the authors threshold to generate ‘hard’ assignments.
  • Regarding the performance of the baseline methods, the authors note that good performance seems to depend critically on predicting hard memberships to multiple circles, using a combination of node and edge information; none of the baselines exhibit precisely this combination, a shortcoming the model addresses.
  • Both of the features the authors propose perform revealing that both schemes encode similar information, which is not surprising, Figure 4: Three detected circles on a small ego-network from Facebook, compared to three groundtruth circles (BER 0.81).
Related work
  • Topic-modeling techniques have been used to uncover ‘mixedmemberships’ of nodes to multiple groups [2], and extensions allow entities to be attributed with text information [3, 5, 11, 13, 26]. Classical algorithms tend to identify communities based on node features [9] or graph structure [1, 21], but rarely use both in concert. Our work is related to [30] in the sense that it performs clustering on social-network data, and [23], which models memberships to multiple communities. Finally, there are works that model network data similar to ours [6, 17], though the underlying models do not form communities. As we shall see, our problem has unique characteristics that require a new model. An extended version of our article appears in [15].
Funding
  • This research has been supported in part by NSF IIS-1016909, CNS-1010921, IIS-1159679, DARPA XDATA, DARPA GRAPHS, Albert Yu & Mary Bechmann Foundation, Boeing, Allyes, Samsung, Intel, Alfred P
Reference
  • Y.-Y. Ahn, J. Bagrow, and S. Lehmann. Link communities reveal multiscale complexity in networks. Nature, 2010.
    Google ScholarLocate open access versionFindings
  • E. Airoldi, D. Blei, S. Fienberg, and E. Xing. Mixed membership stochastic blockmodels. JMLR, 2008.
    Google ScholarLocate open access versionFindings
  • R. Balasubramanyan and W. Cohen. Block-LDA: Jointly modeling entity-annotated text and entity-entity links. In SDM, 2011.
    Google ScholarLocate open access versionFindings
  • E. Boros and P. Hammer. Pseudo-boolean optimization. Discrete Applied Mathematics, 2002.
    Google ScholarLocate open access versionFindings
  • J. Chang and D. Blei. Relational topic models for document networks. In AIStats, 2009.
    Google ScholarLocate open access versionFindings
  • J. Chang, J. Boyd-Graber, and D. Blei. Connections between the lines: augmenting social networks with text. In KDD, 2009.
    Google ScholarLocate open access versionFindings
  • Y. Chen and C. Lin. Combining SVMs with various feature selection strategies. Springer, 2006.
    Google ScholarFindings
  • M. Handcock, A. Raftery, and J. Tantrum. Model-based clustering for social networks. Journal of the Royal Statistical Society Series A, 2007.
    Google ScholarLocate open access versionFindings
  • S. Johnson. Hierarchical clustering schemes. Psychometrika, 1967.
    Google ScholarLocate open access versionFindings
  • D. Kim, Y. Jo, L.-C. Moon, and A. Oh. Analysis of twitter lists as a potential source for discovering latent characteristics of users. In CHI, 2010.
    Google ScholarLocate open access versionFindings
  • P. Krivitsky, M. Handcock, A. Raftery, and P. Hoff. Representing degree distributions, clustering, and homophily in social networks with latent cluster random effects models. Social Networks, 2009.
    Google ScholarLocate open access versionFindings
  • P. Lazarsfeld and R. Merton. Friendship as a social process: A substantive and methodological analysis. In Freedom and Control in Modern Society. 1954.
    Google ScholarLocate open access versionFindings
  • Y. Liu, A. Niculescu-Mizil, and W. Gryc. Topic-link LDA: joint models of topic and author community. In ICML, 2009.
    Google ScholarLocate open access versionFindings
  • D. MacKay. Information Theory, Inference and Learning Algorithms. Cambrdige University Press, 2003.
    Google ScholarFindings
  • J. McAuley and J. Leskovec. Discovering social circles in ego networks. arXiv:1210.8182, 2012.
    Findings
  • M. McPherson. An ecology of affiliation. American Sociological Review, 1983.
    Google ScholarLocate open access versionFindings
  • A. Menon and C. Elkan. Link prediction via matrix factorization. In ECML/PKDD, 2011.
    Google ScholarLocate open access versionFindings
  • A. Mislove, B. Viswanath, K. Gummadi, and P. Druschel. You are who you know: Inferring user profiles in online social networks. In WSDM, 2010.
    Google ScholarLocate open access versionFindings
  • P. Nasirifard and C. Hayes. Tadvise: A twitter assistant based on twitter lists. In SocInfo, 2011.
    Google ScholarLocate open access versionFindings
  • M. Newman. Modularity and community structure in networks. PNAS, 2006.
    Google ScholarLocate open access versionFindings
  • G. Palla, I. Derenyi, I. Farkas, and T. Vicsek. Uncovering the overlapping community structure of complex networks in nature and society. Nature, 2005.
    Google ScholarLocate open access versionFindings
  • C. Rother, V. Kolmogorov, V. Lempitsky, and M. Szummer. Optimizing binary MRFs via extended roof duality. In CVPR, 2007.
    Google ScholarLocate open access versionFindings
  • A. Streich, M. Frank, D. Basin, and J. Buhmann. Multi-assignment clustering for boolean data. JMLR, 2012.
    Google ScholarLocate open access versionFindings
  • J. Ugander, B. Karrer, L. Backstrom, and C. Marlow. The anatomy of the Facebook social graph. preprint, 2011.
    Google ScholarFindings
  • C. Volinsky and A. Raftery. Bayesian information criterion for censored survival models. Biometrics, 2000.
    Google ScholarLocate open access versionFindings
  • D. Vu, A. Asuncion, D. Hunter, and P. Smyth. Dynamic egocentric models for citation networks. In
    Google ScholarLocate open access versionFindings
  • S. Wu, J. Hofman, W. Mason, and D. Watts. Who says what to whom on twitter. In WWW, 2011.
    Google ScholarLocate open access versionFindings
  • J. Yang and J. Leskovec. Community-affiliation graph model for overlapping community detection. In
    Google ScholarLocate open access versionFindings
  • J. Yang and J. Leskovec. Defining and evaluating network communities based on ground-truth. In ICDM, 2012.
    Google ScholarLocate open access versionFindings
  • T. Yoshida. Toward finding hidden communities based on user profiles. In ICDM Workshops, 2010.
    Google ScholarLocate open access versionFindings
  • J. Zhao. Examining the evolution of networks based on lists in twitter. In IMSAA, 2011.
    Google ScholarLocate open access versionFindings
Your rating :
0

 

Tags
Comments
数据免责声明
页面数据均来自互联网公开来源、合作出版商和通过AI技术自动分析结果,我们不对页面数据的有效性、准确性、正确性、可靠性、完整性和及时性做出任何承诺和保证。若有疑问,可以通过电子邮件方式联系我们:report@aminer.cn
小科