2019
DOI: 10.1007/s00778-019-00544-1
|View full text |Cite
|
Sign up to set email alerts
|

Comparing heuristics for graph edit distance computation

Abstract: Because of its flexibility, intuitiveness, and expressivity, the graph edit distance (GED) is one of the most widely used distance measures for labeled graphs. Since exactly computing GED is NP-hard, over the past years, various heuristics have been proposed. They use techniques such as transformations to the linear sum assignment problem with error-correction, local search, and linear programming to approximate GED via upper or lower bounds. In this paper, we provide a systematic overview of the most importan… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1

Citation Types

0
65
0

Year Published

2019
2019
2024
2024

Publication Types

Select...
4
2
1

Relationship

0
7

Authors

Journals

citations
Cited by 47 publications
(65 citation statements)
references
References 66 publications
0
65
0
Order By: Relevance
“…In literature, there are four main graph classification methods: graph isomorphism [17], graph edit distance [18], [19], graph kernels [8], [20]- [25] and graph neural networks [26]- [28]. These methods basically measure the similarity between networks by comparing the three network structures in these networks: network topology, vertex weights/attributes, and edge weights/attributes [1].…”
Section: ) Graph Classificationmentioning
confidence: 99%
See 1 more Smart Citation
“…In literature, there are four main graph classification methods: graph isomorphism [17], graph edit distance [18], [19], graph kernels [8], [20]- [25] and graph neural networks [26]- [28]. These methods basically measure the similarity between networks by comparing the three network structures in these networks: network topology, vertex weights/attributes, and edge weights/attributes [1].…”
Section: ) Graph Classificationmentioning
confidence: 99%
“…While GED is very popular for many graph mining problems, exact GED computation is NP-hard. Therefore, many different heuristic algorithms are proposed to compute approximate solutions [19], [29]- [32]. Among them, local search based algorithms [29], [30] provide the tightest upper bounds for GED.…”
Section: ) Graph Classificationmentioning
confidence: 99%
“…We conduct all experiments in this paper using the PHOCNet implementation 3 by Sudholt and Fink [24]. For the sample selection, we use the IPM implementation 4 by Zaeemzadeh et al [29] and we compute the HED between graphs using the HED implementation in GEDLIB 5 by Blumenthal et al [4,5], which we modified to support the cost model by Stauffer et al [23]. The code used to conduct all experiments as well as the raw data is available online 6 .…”
Section: Experiments Setupmentioning
confidence: 99%
“…114 6. 16 The performance of our method for the one-shot action classification task using the CAD-120 dataset as reported in Tab. 6.4 is further analyzed here.…”
Section: 13mentioning
confidence: 99%
“…Since exactly computing GED is NP-hard, even for uniform edit costs, while in practice, no available exact algorithm can reliably compute GED on graphs with more than 16 nodes, various heuristics have been proposed over the past years [16,66]. Various techniques are employed, such as formulation of the problem to the linear sum assignment problem with error correction, local search, and linear programming to approximate GED via upper or lower bounds.…”
mentioning
confidence: 99%