Dense community detection in multi-valued attributed networks

Huang, Xin; Cheng, Hong; Yu, Jeffrey Xu

doi:10.1016/j.ins.2015.03.075

Cited by 62 publications

(37 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With the increasing dimensionality of attribute space, the discrimination power of the attribute distance or similarity in full space may decrease [8]. Thus attribute subspace community detection methods [22,6,7,8,23,9,24,25,26,27] mining communities with nodes similar in attribute subspaces are preferred in most cases. CoPaM method [22] mines dense and connected subgraphs with homogeneous values in attribute subspaces by efficient pruning strategies.…”

Section: Related Workmentioning

confidence: 99%

“…GAMer method [8] defines twofold clusters by combing the paradigms of dense subgraph mining and attribute subspace clustering and mines them by various pruning strategies. SCMAG method [9] identifies cell-based subspace clusters composed of cells with dense coverage and connectivity in the subspace. Both of attribute full space and subspace methods mentioned above are unsupervised.…”

Section: Related Workmentioning

confidence: 99%

“…Community structure is one of the most prominent features of social networks, as it helps to visualize the network structures [1], enhance the information retrieval and promote the products recommendation [2], etc.. Social networks are typical attributed networks whose nodes are associated with attributes. Different from communities detected only based on structure cohesiveness property [3,4,5], communities with attribute subspaces [6,7,8,9] not only have structure cohesiveness property, but also have attribute similarity property. An attribute subspace is a subset of the whole attribute dimension set.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Mining application-aware community organization with expanded feature subspaces from concerned attributes in social networks

Pan

2018

Knowledge-Based Systems

View full text Add to dashboard Cite

Social networks are typical attributed networks with node attributes. Different from traditional attribute community detection problem aiming at obtaining the whole set of communities in the network, we study an application-oriented problem of mining an application-aware community organization with respect to specific concerned attributes. The concerned attributes are designated based on the requirements of any application by a user in advance. The application-aware community organization w.r.t. concerned attributes consists of the communities with feature subspaces containing these concerned attributes. Besides concerned attributes, feature subspace of each required community may contain some other relevant attributes. All relevant attributes of a feature subspace jointly describe and determine the community embedded in such subspace. Thus the problem includes two subproblems, i.e., how to expand the set of concerned attributes to complete feature subspaces and how to mine the communities embedded in the expanded subspaces. Two subproblems are jointly solved by optimizing a quality function called subspace fitness. An algorithm called ACM is proposed. In order to locate the communities potentially belonging to the applicationaware community organization, cohesive parts of a network backbone composed of nodes with similar concerned attributes are detected and set as the community seeds. The set of concerned attributes is set as the initial subspace for all community seeds. Then each community seed and its attribute subspace are adjusted iteratively to optimize the subspace fitness. Extensive experiments on synthetic datasets demonstrate the effectiveness and efficiency of our method and applications on real-world networks show its application values.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Mining application-aware community organization with expanded feature subspaces from concerned attributes in social networks

Pan

2018

Knowledge-Based Systems

View full text Add to dashboard Cite

show abstract

“…Some other approaches are based in the optimization of local communities [10], but some of them ignores the connection between attributes as ENCLUS [9]. We use CESNA approach to measure the performance of the detection of overlapping communities with and without RELNA in order to evaluate the performance with some standard measures.…”

Section: Overlapping Community Detection In Attributed Networkmentioning

confidence: 99%

“…It has been proved that adding node attributes helps to recover the underlying community structure in content-rich networks more effectively than using links alone [26] . Integration of attribute data becomes a recent challenge of clustering attributed graphs [9][26] [33]. One of the first works that combine node attributes and graph structure was done by Zhou et.…”

Section: Introductionmentioning

confidence: 99%

Ranking features in Facebook to detect overlapping communities

Martinez-Seis

2016

2016 IEEE 13th International Conference on Networking, Sensing, and Control (ICNSC)

View full text Add to dashboard Cite

Attributed social networks includes not only the network structure but also features of the nodes. The huge data in social networks has increased the risk of false discoveries, identifying communities in Facebook requires simple but effective fast techniques. Existing community detection methods suffer from high computational cost caused by both huge structure (nodes) and attribute dimensionality.Based on the homophily property of social networks, that linked nodes likely share similar attributes, we propose a new approach to rank attributes of a social network, called RELNA. Then we detect communities using only high ranked attributes instead of all, therefore, the computation time can be reduced dramatically. Comparisons between RELNA and LINKREC (a well-known attribute ranking algorithm) were made through experiments on Facebook data sets, the results are almost the same. Furthermore, we use the obtained top relevant attributes on CESNA algorithm (an well known overlapping community detection algorithm) to detect communities, results show that our approach is much faster meanwhile cluster precision keeps almost the same as CESNA.

show abstract