@inproceedings{xiang2010modeling,
  0 = {http://portal.acm.org/citation.cfm?id=1772690.1772790},
  1 = {http://dx.doi.org/10.1145/1772690.1772790},
  abstract = {Previous work analyzing social networks has mainly focused on binary friendship relations. However, in online social networks the low cost of link formation can lead to networks with heterogeneous relationship strengths (e.g., acquaintances and best friends mixed together). In this case, the binary friendship indicator provides only a coarse representation of relationship information. In this work, we develop an unsupervised model to estimate relationship strength from interaction activity (e.g., communication, tagging) and user similarity. More specifically, we formulate a link-based latent variable model, along with a coordinate ascent optimization procedure for the inference. We evaluate our approach on real-world data from Facebook and LinkedIn, showing that the estimated link weights result in higher autocorrelation and lead to improved classification accuracy.},
  address = {New York, NY, USA},
  at = {2010-04-28 16:48:39},
  author = {Xiang, Rongjing and Neville, Jennifer and Rogati, Monica},
  booktitle = {WWW '10: Proceedings of the 19th international conference on World wide web},
  doi = {10.1145/1772690.1772790},
  id = {7097815},
  interhash = {8b1a72c3e48b8327d0494742eb6a8a43},
  intrahash = {e7f26fd76f6e97b021c33ce942cea3ef},
  isbn = {978-1-60558-799-8},
  location = {Raleigh, NC, USA},
  month = {April},
  pages = {981--990},
  priority = {2},
  publisher = {ACM},
  title = {Modeling relationship strength in online social networks},
  url = {http://dx.doi.org/10.1145/1772690.1772790},
  year = 2010
}

@misc{leskovec2010empirical,
  abstract = {  Detecting clusters or communities in large real-world graphs such as largesocial or information networks is a problem of considerable interest. Inpractice, one typically chooses an objective function that captures theintuition of a network cluster as set of nodes with better internalconnectivity than external connectivity, and then one applies approximationalgorithms or heuristics to extract sets of nodes that are related to theobjective function and that "look like" good communities for the application ofinterest. In this paper, we explore a range of network community detectionmethods in order to compare them and to understand their relative performanceand the systematic biases in the clusters they identify. We evaluate severalcommon objective functions that are used to formalize the notion of a networkcommunity, and we examine several different classes of approximation algorithmsthat aim to optimize such objective functions. In addition, rather than simplyfixing an objective and asking for an approximation to the best cluster of anysize, we consider a size-resolved version of the optimization problem.Considering community quality as a function of its size provides a much finerlens with which to examine community detection algorithms, since objectivefunctions and approximation algorithms often have non-obvious size-dependentbehavior.},
  author = {Leskovec, Jure and Lang, Kevin J. and Mahoney, Michael W.},
  interhash = {0e58de655596b2198f4a7001facd0c32},
  intrahash = {410a9cbea51ea5dd3c56aad26a0e11b2},
  note = {cite arxiv:1004.3539},
  title = {Empirical Comparison of Algorithms for Network Community Detection},
  url = {http://arxiv.org/abs/1004.3539},
  year = 2010
}

@misc{leskovec2010predicting,
  abstract = {  We study online social networks in which relationships can be either positive(indicating relations such as friendship) or negative (indicating relationssuch as opposition or antagonism). Such a mix of positive and negative linksarise in a variety of online settings; we study datasets from Epinions,Slashdot and Wikipedia. We find that the signs of links in the underlyingsocial networks can be predicted with high accuracy, using models thatgeneralize across this diverse range of sites. These models provide insightinto some of the fundamental principles that drive the formation of signedlinks in networks, shedding light on theories of balance and status from socialpsychology; they also suggest social computing applications by which theattitude of one user toward another can be estimated from evidence provided bytheir relationships with other members of the surrounding social network.},
  author = {Leskovec, Jure and Huttenlocher, Daniel and Kleinberg, Jon},
  interhash = {24158224b6b45342017e1157f98f5c65},
  intrahash = {cdf322be85a607c789a5ee0e930f72ef},
  note = {cite arxiv:1003.2429},
  title = {Predicting Positive and Negative Links in Online Social Networks},
  url = {http://arxiv.org/abs/1003.2429},
  year = 2010
}