Communities validity: methodical evaluation of community mining algorithms

Rabbany, Reihaneh; Takaffoli, Mansoureh; Fagnan, Justin; Zaı̈ane, Osmar R.; Campello, Ricardo J. G. B.

doi:10.1007/s13278-013-0132-x

Cited by 11 publications

(22 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We, thus, suggest to adj ust the NMI with respect to a new property, called the reference closeness property, that is, the closer the number of communities found by an algorithm to the number of clusters of the reference clustering, the higher the MI value should be. This desirable behavior of any evaluation criterion has been pointed out by Rabbany et al (2013), where they state that the "ideal behaviour of an index should be that it gives low scores for partitionings/fragmentations in which the number of clusters is much higher or lower than what we have in a ground-truth." Moreover, Vinh et al (2010) experimentally found that, in the context of consensus clustering Strehl and Ghosh (2002), a clustering having a number of clusters coincident with the true cluster number is more robust.…”

Section: Adjustment For Closenessmentioning

confidence: 95%

“…These indices have often been borrowed and modified to evaluate the many methods proposed for community mining. Rabbany et al (), especially, considered the same classification scheme for community detection methods, investigated quality criteria for internal, relative, and external evaluation and modified some validity indices to make them apt for network data.…”

Section: Definitions and Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Correction for Closeness: Adjusting Normalized Mutual Information Measure for Clustering Comparison

Amelio

Pizzuti

2016

Computational Intelligence

View full text Add to dashboard Cite

Normalized mutual information (NMI) is a widely used measure to compare community detection methods. Recently, however, the need of adjustment for information theory-based measures has been argued because of the so-called selection bias problem, that is, they show the tendency in choosing clustering solutions with more communities. In this article, an experimental evaluation of these measures is performed to deeply investigate the problem, and an adjustment that scales the values of these measures is proposed. Experiments on synthetic networks, for which the ground-truth division is known, highlight that scaled NMI does not present the selection bias behavior. Moreover, a comparison among some well-known community detection methods on synthetic generated networks shows a fairer behavior of scaled NMI, especially when the network topology does not present a clear community structure. The experimentation also on two real-world networks reveals that the corrected formula allows to choose, among a set, the method finding a network division that better reflects the ground-truth structure.

show abstract

Section: Adjustment For Closenessmentioning

confidence: 95%

Section: Definitions and Related Workmentioning

confidence: 99%

Correction for Closeness: Adjusting Normalized Mutual Information Measure for Clustering Comparison

Amelio

Pizzuti

2016

Computational Intelligence

View full text Add to dashboard Cite

show abstract

“…The Constant Baseline [3], [5], [10], [13], [20] property deals with statistical independence, i.e. the case where one compares two partitions sampled independently at random.…”

Section: ) Handling Of Independent Partitionsmentioning

confidence: 99%

“…This is why the second approach, which is empirical, is much more frequent in the literature (e.g. [7], [10]). It consists in applying some predefined transformations to certain partitions, both designed in a way that is related to the property of interest, and to study how the measure reacts to these perturbations by using it to compare those partitions.…”

Section: Introductionmentioning

confidence: 99%

Characterizing and Comparing External Measures for the Assessment of Cluster Analysis and Community Detection

2021

View full text Add to dashboard Cite

In the context of cluster analysis and graph partitioning, many external evaluation measures have been proposed in the literature to compare two partitions of the same set. This makes the task of selecting the most appropriate measure for a given situation a challenge for the end user. However, this issue is overlooked in the literature. Researchers tend to follow tradition and use the standard measures of their field, although they often became standard only because previous researchers started consistently using them. In this work, we propose a new empirical evaluation framework to solve this issue, and help the end user selecting an appropriate measure for their application. For a collection of candidate measures, it first consists in describing their behavior by computing them for a generated dataset of partitions, obtained by applying a set of predefined parametric partition transformations. Second, our framework performs a regression analysis to characterize the measures in terms of how they are affected by these parameters and transformations. This allows both describing and comparing the measures. Our approach is not tied to any specific measure or application, so it can be applied to any situation. We illustrate its relevance by applying it to a selection of standard measures, and show how it can be put in practice through two concrete use cases.

show abstract

“…8 Community structure refers to a set of vertices whose edges inside are more dense than edges outside. 9 Finding community in a network known as a community detection problem (also called community structure identification or cluster finding). Community detection plays a significant role for studying and understanding the structure and function of real-world networks including several various domains such as clustering web stations having similar functionality and are geographically near to each other may improve the performance of services provided on the World Wide Web.…”

Section: Introductionmentioning

confidence: 99%

Distributed learning automata-based algorithm for community detection in complex networks

Khomami

Rezvanian

Meybodi

2016

Int. J. Mod. Phys. B

View full text Add to dashboard Cite

Community structure is an important and universal topological property of many complex networks such as social and information networks. The detection of communities of a network is a significant technique for understanding the structure and function of networks. In this paper, we propose an algorithm based on distributed learning automata for community detection (DLACD) in complex networks. In the proposed algorithm, each vertex of network is equipped with a learning automation. According to the cooperation among network of learning automata and updating action probabilities of each automaton, the algorithm interactively tries to identify high-density local communities. The performance of the proposed algorithm is investigated through a number of simulations on popular synthetic and real networks. Experimental results in comparison with popular community detection algorithms such as walk trap, Danon greedy optimization, Fuzzy community detection, Multi-resolution community detection and label propagation demonstrated the superiority of DLACD in terms of modularity, NMI, performance, min-max-cut and coverage.

show abstract

Communities validity: methodical evaluation of community mining algorithms

Cited by 11 publications

References 50 publications

Correction for Closeness: Adjusting Normalized Mutual Information Measure for Clustering Comparison

Correction for Closeness: Adjusting Normalized Mutual Information Measure for Clustering Comparison

Characterizing and Comparing External Measures for the Assessment of Cluster Analysis and Community Detection

Distributed learning automata-based algorithm for community detection in complex networks

Contact Info

Product

Resources

About