Empirical evaluation of applying ensemble methods to ego-centred community identification in complex networks

Kanawati, Rushed

doi:10.1016/j.neucom.2014.09.042

Cited by 19 publications

(11 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Another form to return a single partition from a given set of solutions is by consensus clustering strategies. Kanawati (2015) suggested using different consensus and ensemble strategies to obtain a partition from outputs of a graph clustering algorithm. Nevertheless, the method of Kanawati (2015) was designed only to find clusters of target nodes in a distributed form.…”

Section: Solution Selection For the Decision-makingmentioning

confidence: 99%

“…Kanawati (2015) suggested using different consensus and ensemble strategies to obtain a partition from outputs of a graph clustering algorithm. Nevertheless, the method of Kanawati (2015) was designed only to find clusters of target nodes in a distributed form. In this paper, we propose a consensus strategy to define a partition from the Pareto solutions, instead of employing the measure-based strategies introduced in the literature that are biased to a single evaluation metric.…”

Section: Solution Selection For the Decision-makingmentioning

confidence: 99%

“…Liang et al (2014) combined a consensus strategy with a label propagation (LP) algorithm (Raghavan et al, 2007) to obtain better partitions than LP. As previously mentioned, although Kanawati (2015) approached a graph clustering problem which does not find a partitioning, one of the strategies the author employed is founded on the definition of a consensus matrix, similar to the strategy that we introduce in this paper.…”

Section: Consensus Clusteringmentioning

confidence: 99%

See 2 more Smart Citations

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Tautenhain

Nascimento

2020

Expert Systems with Applications

View full text Add to dashboard Cite

Graph clustering is a challenging pattern recognition problem whose goal is to identify vertex partitions with high intra-group connectivity. This paper investigates a bi-objective problem that maximizes the number of intra-cluster edges of a graph and minimizes the expected number of inter-cluster edges in a random graph with the same degree sequence as the original one. The difference between the two investigated objectives is the definition of the well-known measure of graph clustering quality: the modularity. We introduce a spectral decomposition hybridized with an evolutionary heuristic, called MOSpecG, to approach this bi-objective problem and an ensemble strategy to consolidate the solutions found by MOSpecG into a final robust partition. The results of computational experiments with real and artificial LFR networks demonstrated a significant improvement in the results and performance of the introduced method in regard to another bi-objective algorithm found in the literature. The crossover operator based on the geometric interpretation of the modularity maximization problem to match the communities of a pair of individuals was of utmost importance for the good performance of MOSpecG. Hybridizing spectral graph theory and intelligent systems allowed us to define significantly high-quality community structures.

show abstract

Section: Solution Selection For the Decision-makingmentioning

confidence: 99%

Section: Solution Selection For the Decision-makingmentioning

confidence: 99%

Section: Consensus Clusteringmentioning

confidence: 99%

See 1 more Smart Citation

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Tautenhain

Nascimento

2020

Expert Systems with Applications

View full text Add to dashboard Cite

show abstract

“…The approach presented by Zakrzewska and Bader (2015) employs a seed set expansion procedure that incrementally updates the community as the underlying graph changes. Kanawati (2015) evaluates the impact of applying different ways of combining multiple local community functions to identify the node-centric communities. The Lemon algorithm proposed by Li et al (2015) exploits truncated random walks and approximate invariant subspace to discover a local community for any given seed set.…”

Section: Related Workmentioning

confidence: 99%

Local community detection in multilayer networks

Interdonato

Tagarelli

Ienco³

et al. 2017

Data Min Knowl Disc

View full text Add to dashboard Cite

The problem of local community detection in graphs refers to the identification of a community that is specific to a query node and relies on limited information about the network structure. Existing approaches for this problem are defined to work in dynamic network scenarios, however they are not designed to deal with complex real-world networks, in which multiple types of connectivity might be considered. In this work, we fill this gap in the literature by introducing the first framework for local community detection in multilayer networks (ML-LCD). We formalize the ML-LCD optimization problem and provide three definitions of the associated objective function, which correspond to different ways to incorporate within-layer and across-layer topological features. We also exploit our framework to generate multilayer global community structures. We conduct an extensive experimentation using seven real-world multilayer networks, which also includes comparison with state-ofthe-art methods for single-layer local community detection and for multilayer global community detection. Results show the significance of our proposed methods in discovering local communities over multiple layers, and also highlight their ability in producing global community structures that are better in modularity than those produced by native global community detection approaches.

show abstract

“…It is the case of node classifiers based on random walks, which propagate the node features through the graph [67,81]. A special case of node classification is explored in a work from Kanawati [45], where communities are not detected across the graph but rather predicted for single nodes.…”

Section: Social Classificationmentioning

confidence: 99%

Beyond social graphs: mining patterns underlying social interactions

Baldominos

Calle

Cuadra

2016

Pattern Anal Applic

View full text Add to dashboard Cite

This work aims at discovering and extracting relevant patterns underlying social interactions. To do so, some knowledge extracted from Facebook, a social networking site, is formalised by means of an Extended Social Graph, a data structure which goes beyond the original concept of a social graph by also incorporating information on interests. When the Extended Social Graph is built, state-of-the-art techniques are applied over it in order to discover communities. Once these social communities are found, statistical techniques will look for relevant patterns common to each of those, in such a way that each cluster of users is characterised by a set of common features. The resulting knowledge will be used to develop and evaluate a social recommender system, which aims at suggesting users in a social network with possible friends or interests.

show abstract

Empirical evaluation of applying ensemble methods to ego-centred community identification in complex networks

Cited by 19 publications

References 31 publications

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

An ensemble based on a bi-objective evolutionary spectral algorithm for graph clustering

Local community detection in multilayer networks

Beyond social graphs: mining patterns underlying social interactions

Contact Info

Product

Resources

About