Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2506.20111 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915358221271040 |
|---|---|
| author | Guarracino, Mario R. Miasnikof, Pierre Shestopaloff, Alexander Y. Demni, Houyem Bravo, Cristián Lawryshyn, Yuri |
| author_facet | Guarracino, Mario R. Miasnikof, Pierre Shestopaloff, Alexander Y. Demni, Houyem Bravo, Cristián Lawryshyn, Yuri |
| contents | In this article, we extend a statistical test of graph clusterability, the $δ$ test, to directed graphs with no self loops. The $δ$ test, originally designed for undirected graphs, is based on the premise that graphs with a clustered structure display a mean local density that is statistically higher than the graph's global density. We posit that graphs that do not meet this necessary (but not sufficient) condition for clusterability can be considered unsuited to clustering. In such cases, vertex clusters do not offer a meaningful summary of the broader graph. Additionally in this study, we aim to determine the optimal sample size (number of neighborhoods). Our test, designed for the analysis of large networks, is based on sampling subsets of neighborhoods/nodes. It is designed for cases where computing the density of every node's neighborhood is infeasible. Our results show that the $δ$ test performs very well, even with very small samples of neighborhoods ($1\%$). It accurately detects unclusterable graphs and is also shown to be robust to departures from the underlying assumptions of the $t$ test. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_20111 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A clusterability test for directed graphs Guarracino, Mario R. Miasnikof, Pierre Shestopaloff, Alexander Y. Demni, Houyem Bravo, Cristián Lawryshyn, Yuri Networking and Internet Architecture 62 In this article, we extend a statistical test of graph clusterability, the $δ$ test, to directed graphs with no self loops. The $δ$ test, originally designed for undirected graphs, is based on the premise that graphs with a clustered structure display a mean local density that is statistically higher than the graph's global density. We posit that graphs that do not meet this necessary (but not sufficient) condition for clusterability can be considered unsuited to clustering. In such cases, vertex clusters do not offer a meaningful summary of the broader graph. Additionally in this study, we aim to determine the optimal sample size (number of neighborhoods). Our test, designed for the analysis of large networks, is based on sampling subsets of neighborhoods/nodes. It is designed for cases where computing the density of every node's neighborhood is infeasible. Our results show that the $δ$ test performs very well, even with very small samples of neighborhoods ($1\%$). It accurately detects unclusterable graphs and is also shown to be robust to departures from the underlying assumptions of the $t$ test. |
| title | A clusterability test for directed graphs |
| topic | Networking and Internet Architecture 62 |
| url | https://arxiv.org/abs/2506.20111 |