From Dense Graphs to Meaningful Communities: Assessing Community Quality Using Geodesic Distance Modularity on Metric Backbone-Sparsified Networks
Bibliographic record
Abstract
Community detection serves as a key component of network science, essential for revealing and examining structures within complex networks. This study offers a thorough examination that combines theoretical concepts with detailed practical evaluation. At the heart of our investigation are two interrelated methodologies: the Geodesic Distance Metric (GDM), which serves as a structure-aware, label-independent evaluation tool; and the Metric Backbone, a parameter-free sparsification approach that maintains essential community structures. We present and assess a scalable pipeline for community detection that utilizes Metric Backbone sparsification in conjunction with community detection algorithms, with outcomes evaluated through various well-known quality metrics. Our systematic empirical evaluation across seven diverse real-world datasets shows that Metric Backbone sparsification leads to significant edge reduction-ranging from $14 \%$ in sparse citation networks to $71 \%$ in dense social networks-while preserving community detection quality. Interestingly, the High School dataset maintains F1 and NMI scores of 0.970 even with a $71 \%$ reduction in edges, whereas the Amazon co-purchase network demonstrates consistent performance with a $29 \%$ sparsification. The results emphasize that Metric Backbone sparsification improves computational efficiency and facilitates the use of geodesic-based metrics such as GDM in identifying compact, well-separated communities. This thorough analysis presents the initial systematic review of geodesic-based community evaluation on structurepreserving sparsified networks, delivering valuable insights into the conditions under which parameter-free sparsification preserves rather than obstructs community detection across various network types.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.005 | 0.044 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.006 | 0.003 |
| Science and technology studies | 0.001 | 0.002 |
| Scholarly communication | 0.002 | 0.005 |
| Open science | 0.001 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".