| Title: | Graphlet-based edge weighting for improved community detection in complex networks |
|---|
| Authors: | ID Anastasiia, Dziuba (Author)
ID Pražnikar, Jure (Author) |
|---|
| Files: |  RAZ_Anastasiia_Dziuba_2026.pdf (6,14 MB)
MD5: F489059387B3816AB9D98C2D37C254F4
 https://link.springer.com/article/10.1007/s10844-026-01073-4
|
|---|
| Language: | English |
|---|
| Work type: | Article |
|---|
| Typology: | 1.01 - Original Scientific Article |
|---|
| Organization: | IAM - Andrej Marušič Institute
|
|---|
| Abstract: | Community detection reveals the functional organization of complex networks. Unlike edge-based methods, graphlet-based approaches use higher-order structures, but most studies focus on common motifs or lack validation on networks with known communities. We evaluate eight small motifs on synthetic and real-world networks and propose a framework that weights node pairs by their graphlet co-occurrence while preserving original edges. This captures both direct connections and local topology, improving community detection. Our results show that graphlet-based weighting consistently enhances performance, but no single motif is best for all networks. Simpler motifs can perform as well as dense cliques, suggesting that relying only on cliques may miss important connectivity patterns. |
|---|
| Keywords: | community, network, graphlets, small-motif |
|---|
| Publication version: | Author Accepted Manuscript |
|---|
| Publication date: | 01.07.2026 |
|---|
| Year of publishing: | 2026 |
|---|
| Number of pages: | str. 1-24 |
|---|
| Numbering: | Vol. , iss. |
|---|
| PID: | 20.500.12556/RUP-23204  |
|---|
| UDC: | 004 |
|---|
| ISSN on article: | 0925-9902 |
|---|
| DOI: | 10.1007/s10844-026-01073-4 |
|---|
| COBISS.SI-ID: | 283297539 |
|---|
| Publication date in RUP: | 01.07.2026 |
|---|
| Views: | 30 |
|---|
| Downloads: | 2 |
|---|
| Metadata: |    |
|---|
|
:
|
Copy citation |
|---|
| | | | Average score: | (0 votes) |
|---|
| Your score: | Voting is allowed only for logged in users. |
|---|
| Share: |  |
|---|
Hover the mouse pointer over a document title to show the abstract or click
on the title to get all document metadata. |
