Comparison of Graph Node Distances on Clustering Tasks

Sommer, Felix; Fouss, François; Saerens, Marco

doi:10.1007/978-3-319-44778-0_23

Cited by 26 publications

(32 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparing the results in Table 1 to those in [43] shows that, as expected, the lack of knowledge of the true number of clusters has a detrimental effect on clustering quality, which also cannot be compensated by optimizing kernel parameters per dataset. This behavior should however be expected, as thus far no method has been found to deliver consistently good clustering results, especially with limited to no knowledge on the data.…”

Section: Resultsmentioning

confidence: 58%

“…Varying the graph node distance between a graph's nodes can have a substantial influence on clustering effectiveness. As shown in prior work [26,43], modern graph distances can improve clustering solutions. However, clustering the nodes using a kernel k-means approach still suffers a major drawback, in that the k-means algorithm always requires that the number of clusters is given.…”

Section: Introductionmentioning

confidence: 91%

“…Due to space restrictions, for the following four distances and (dis-)similarities already investigated in our previous work [43], we refer to this paper (Eq. (1) through (4) in Sect.…”

Section: Background and Notationmentioning

confidence: 99%

“…This extended method optimizes kernel parameters and automatically estimates the natural number of clusters present in the dataset. The kernel k-means algorithm itself is the same as in our previous work, see [43], and corresponds to a two-step iterative algorithm based on a distance-or dissimilarity-matrix instead of features. Given a meaningful, symmetric distance matrix Δ, containing the distances Δ ij , the goal is to partition the nodes by minimizing the total within-cluster sum of squared distances.…”

Section: Kernel K-means Coupled With Modularity Criterionmentioning

confidence: 99%

“…In this work we therefore (i) propose automatic kernel parameter tuning, in order to improve clustering quality, without requiring user interaction, using the modularity criterion, (ii) identify the natural number of clusters present in a dataset, using also the modularity criterion, without knowledge of the (correct) number of clusters beforehand, (iii) finally, compare this improved kernel k-means approach, using the kernels presented in our previous work [43], as well as additional kernels [26], to the Louvain method [2].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Modularity-Driven Kernel k-means for Community Detection

Sommer

Fouss

Saerens

2017

Artificial Neural Networks and Machine Learning – ICANN 2017

Self Cite

View full text Add to dashboard Cite

The k-means algorithm is probably the most well-known and most popular clustering method in existence today. This work evaluates if a new, autonomous, kernel k-means approach for graph node clustering coupled with the modularity criterion can rival, e.g., the well-established Louvain method. We test the algorithm on social network datasets of various sizes and types. The new method estimates the optimal kernel or distance parameters as well as the natural number of clusters in the dataset based on modularity. Results indicate that this simple black-box algorithm manages to perform on par with the Louvain method given the same input. Abstract. The k-means algorithm is probably the most well-known and most popular clustering method in existence today. This work evaluates if a new, autonomous, kernel k-means approach for graph node clustering coupled with the modularity criterion can rival, e.g., the well-established Louvain method. We test the algorithm on social network datasets of various sizes and types. The new method estimates the optimal kernel or distance parameters as well as the natural number of clusters in the dataset based on modularity. Results indicate that this simple black-box algorithm manages to perform on par with the Louvain method given the same input.

show abstract

Section: Resultsmentioning

confidence: 58%

Section: Introductionmentioning

confidence: 91%

“…Due to space restrictions, for the following four distances and (dis-)similarities already investigated in our previous work [43], we refer to this paper (Eq. (1) through (4) in Sect.…”

Section: Background and Notationmentioning

confidence: 99%

Section: Kernel K-means Coupled With Modularity Criterionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Modularity-Driven Kernel k-means for Community Detection

Sommer

Fouss

Saerens

2017

Artificial Neural Networks and Machine Learning – ICANN 2017

Self Cite

View full text Add to dashboard Cite

show abstract

Efficiency of Transformations of Proximity Measures for Graph Clustering

Aynulin

2019

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Graph-Based Fraud Detection with the Free Energy Distance

Courtain

Lebichot

Kivimäki

et al. 2019

Complex Networks and Their Applications VIII

Self Cite

View full text Add to dashboard Cite

This paper investigates a real-world application of the free energy distance between nodes of a graph [14, 20] by proposing an improved extension of the existing Fraud Detection System named APATE [36]. It relies on a new way of computing the free energy distance based on paths of increasing length, and scaling on large, sparse, graphs. This new approach is assessed on a real-world large-scale e-commerce payment transactions dataset obtained from a major Belgian credit card issuer. Our results show that the free-energy based approach reduces the computation time by one half while maintaining state-of-the art performance in term of Precision@100 on fraudulent card prediction.

show abstract

Comparison of Graph Node Distances on Clustering Tasks

Cited by 26 publications

References 20 publications

Modularity-Driven Kernel k-means for Community Detection

Modularity-Driven Kernel k-means for Community Detection

Efficiency of Transformations of Proximity Measures for Graph Clustering

Graph-Based Fraud Detection with the Free Energy Distance

Contact Info

Product

Resources

About