Editorial Trajectories in Wikipedia Reflect Underlying Hyperlink Structure

Hang-Hyun Jo; Mi Jin Lee; Seung-Woo Son; Yeonji Seo; Yohsuke Murase

arxiv: 2605.16850 · v1 · pith:PGWJXUI3new · submitted 2026-05-16 · ⚛️ physics.soc-ph

Editorial Trajectories in Wikipedia Reflect Underlying Hyperlink Structure

Yeonji Seo , Mi Jin Lee , Seung-Woo Son , Hang-Hyun Jo , Yohsuke Murase This is my paper

Pith reviewed 2026-05-19 19:34 UTC · model grok-4.3

classification ⚛️ physics.soc-ph

keywords Wikipediahyperlinkseditorial transitionsinter-event timesJaccard similaritytopical communitieseditor typesnetwork structure

0 comments

The pith

Wikipedia hyperlinks organize the sequence and timing of edits by contributors.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper combines the hyperlink network among English Wikipedia articles with detailed editing histories to test whether the static link structure shapes how editors move between pages. Transitions between hyperlinked article pairs occur with shorter gaps in time than transitions between non-linked pairs. When articles are grouped into 19 topical communities, editors fall into three groups based on how closely their personal sequence of edits matches the links: specialists stay within narrow topics and follow the hyperlink patterns closely, generalists range across many topics with weaker alignment, and bots show the shortest times but low similarity to the link structure. These patterns indicate that the hyperlink network influences the order of editorial activity in addition to guiding readers.

Core claim

Transitions between hyperlinked article pairs have shorter inter-event times than those between non-hyperlinked pairs. Editors' transition networks, when compared to the corresponding hyperlink subnetworks via Jaccard similarity, reveal three distinct types: specialists with low topical diversity, shorter mean inter-event times, and higher Jaccard similarity; generalists with high topical diversity, longer mean inter-event times, and lower Jaccard similarity; and bots with the shortest mean inter-event times yet low Jaccard similarity despite often high topical diversity. The hyperlink structure is thereby linked to the sequential organization of editorial activity.

What carries the argument

Jaccard similarity between each editor's transition network and the hyperlink subnetwork, combined with inter-event time measurements and coarse-graining of the network into 19 topical communities to quantify diversity.

If this is right

Specialist editors remain within limited topical domains and show transition patterns that align closely with the hyperlink structure.
Generalist editors range across broader topics and exhibit weaker similarity between their edit sequences and the hyperlink network.
Bots produce the shortest inter-edit times but the lowest alignment with hyperlinks, separating them from both specialist and generalist human patterns.
The hyperlink structure participates in the sequential organization of edits within collaborative knowledge systems.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar alignments between link structure and contribution sequences may appear in other collaborative platforms that combine navigation links with edit histories.
Platform designers could leverage existing hyperlinks to surface suggested next articles for editors based on observed transition patterns.
Controlled addition or removal of hyperlinks followed by observation of subsequent edit sequences would test whether the network actively directs editing order.

Load-bearing premise

That shorter inter-event times and higher Jaccard similarity between editor transitions and hyperlinks reflect a structural influence of the link network rather than confounding factors such as article popularity or independent editor preferences.

What would settle it

Measuring inter-event times after matching hyperlinked and non-hyperlinked article pairs for comparable edit volume or popularity, or checking whether randomly shuffled transition sequences achieve similar Jaccard similarity to the actual hyperlink network.

Figures

Figures reproduced from arXiv: 2605.16850 by Hang-Hyun Jo, Mi Jin Lee, Seung-Woo Son, Yeonji Seo, Yohsuke Murase.

**Figure 2.** Figure 2: Distributions of IETs for different transition types. (a) Distribution of long-term IETs, P(τlong). The red solid and blue dashed lines represent hyperlinked and non-hyperlinked pairs, respectively. The gray vertical dotted lines indicate characteristic peaks at 484 s (≈ 8 min), 81100 s (≈ 1 day), 170000 s (≈ 2 days), and 259000 s (≈ 3 days). The 484 s peak is a 2002 ‘Conversion script’ artifact, unrelated… view at source ↗

**Figure 3.** Figure 3: Visualization of the characteristics of the detected communities in the Wikipedia hyperlink network. (a) Treemap showing the distribution of articles (community size) across communities. (b) Scatter plot of community-level properties. The 19 distinct markers and colors correspond to the communities detailed in [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗

**Figure 4.** Figure 4: Community-level diversity of editing patterns for individual editors. (a, b) Scatter plots of editor activity, defined as the number of distinct articles edited by an editor, versus diversity, measured by (a) entropy H and (b) the inverse Simpson index 1/λ. Each point corresponds to an editor and is colored according to the number of participating communities. Black-edged squares in (a)-(b) indicate editin… view at source ↗

**Figure 5.** Figure 5: Temporal-structural relationships of editorial behavior across editor types. (a) Violin plots of the Jaccard similarity J by editor type. Each point represents an individual editor. The yellow box indicates the mean, and the vertical line represents the median. (b) Distribution of the mean long-term IET per editor, τ¯long. The dashed lines indicate exponential tail fits of the form exp(−τ¯long/b), from whi… view at source ↗

read the original abstract

Wikipedia hyperlinks have primarily been studied as navigational tools for readers, but their role in how information providers move between articles during editing remains less explored. Here, we combine the hyperlink network among English Wikipedia articles with editorial histories to examine how article-to-article structure is associated with editors' transitions between articles. We first address the temporal aspect of edit transitions by showing that transitions between hyperlinked article pairs have shorter inter-event times (IETs) than those between non-hyperlinked pairs, indicating that connected articles are effectively closer in editing sequences. We then turn to the structural organization of editing behavior by coarse-graining the hyperlink network into 19 topical communities and measuring editors' topical diversity. Finally, we bring the temporal and structural views together by comparing each editor's transition network with the corresponding hyperlink subnetwork using Jaccard similarity. Combining the measures allows us to distinguish three editor types: 'Specialists' are characterized by focused editing within limited topical domains and transition patterns more closely aligned with the hyperlink structure (low topical diversity, shorter mean IETs, and higher Jaccard similarity), whereas 'generalists' cover broader topics and show weaker similarity to the hyperlink structure (high topical diversity, longer mean IETs, and lower Jaccard similarity). 'Bots' show a distinct algorithm-driven behavior, with low Jaccard similarity and the shortest mean IETs, a combination departing from human-editor patterns despite their often high topical diversity. Such findings demonstrate that the hyperlink structure is not just a static scaffold for reader navigation, but is observationally linked to the sequential organization of editorial activity in collaborative knowledge systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The paper links Wikipedia edit sequences to the hyperlink graph via IETs and Jaccard overlap but leaves popularity as a live alternative explanation.

read the letter

The main takeaway is that editors move between Wikipedia articles faster when the pages are already hyperlinked, and their overall transition patterns match the link structure enough to sort them into three groups: specialists who stay close to the links with short times and high overlap, generalists who range across topics with weaker matches, and bots that edit very quickly but show low structural alignment. They get there by measuring inter-event times on linked versus unlinked pairs, detecting 19 topical communities from the hyperlink network, and then comparing each editor's transition network to the corresponding link subnetwork with Jaccard similarity. That combination produces the three-type split and shows the link graph is tied to editing order in the data. The measures themselves are straightforward and pulled directly from the edit logs and the graph, with no obvious self-referential definitions. The work is observational and grounded in a large real-world dataset, which gives it some weight for people studying collaborative platforms. The soft spot is the missing check on article popularity. Popular pages tend to have both higher degree and higher edit volume, so sequences could cluster on them for volume reasons alone rather than because of the specific links. The abstract does not mention degree-matched null models, edit-frequency controls, or rewiring tests that preserve popularity while breaking the hyperlink structure, so the reported differences are still consistent with that simpler account. If the full methods section has those controls, the structural claim strengthens; otherwise it needs them. This is the sort of paper that would interest a reading group working on network effects in online collaboration or digital knowledge production. Readers focused on Wikipedia studies or computational social science could extract usable patterns from the classification, even if the causal story stays tentative. It deserves peer review because the core observations are concrete and the data integration is new enough to be worth referee time, though any review should press hard on robustness against popularity confounds.

Referee Report

2 major / 2 minor

Summary. The paper combines the English Wikipedia hyperlink network with editorial edit histories to investigate associations between article-to-article hyperlink structure and editors' sequential transitions. It reports shorter inter-event times (IETs) for hyperlinked versus non-hyperlinked article pairs, coarse-grains the network into 19 topical communities to assess editors' topical diversity, and uses Jaccard similarity to compare each editor's transition network against the corresponding hyperlink subnetwork. These measures are combined to distinguish three editor types: specialists (low diversity, short IETs, high Jaccard), generalists (high diversity, long IETs, low Jaccard), and bots (short IETs but low Jaccard despite high diversity). The central claim is that hyperlink structure is observationally linked to the organization of editorial activity beyond its role in reader navigation.

Significance. If the reported patterns survive controls for article popularity and edit volume, the work would offer a useful observational bridge between static hyperlink networks and dynamic editing sequences in collaborative knowledge platforms. The integration of temporal (IET), community, and similarity measures to classify editor behaviors is a constructive approach that could inform studies of information production and platform design. The explicit separation of bot behavior from human patterns is a further strength.

major comments (2)

[Abstract and temporal analysis] Abstract and temporal analysis section: the reported shorter IETs for transitions between hyperlinked article pairs versus non-hyperlinked pairs do not include controls for article popularity or edit frequency. Popular articles receive more edits and are more likely to be hyperlinked, so the IET difference could arise from a popularity-driven null rather than hyperlink structure itself. A degree-preserving rewiring null model or regression adjustment for edit volume is needed to support the structural claim.
[Jaccard similarity analysis] Jaccard similarity analysis: the higher Jaccard overlap between specialists' transition networks and hyperlink subnetworks may likewise be driven by the correlation between hyperlink degree and edit volume. Without explicit degree-matching or null models that rewire hyperlinks while preserving the degree sequence, the measure does not isolate the influence of the underlying link structure from confounding editor-article popularity effects.

minor comments (2)

[Community detection] The determination of exactly 19 topical communities should be accompanied by the community detection algorithm employed, its resolution parameter if applicable, and a brief sensitivity check to the number of communities.
[Methods] Clarify the sampling procedure for non-hyperlinked article pairs when computing IETs and the precise definition of an editor's transition network used for the Jaccard calculation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and insightful comments on our manuscript. The concerns regarding potential confounding by article popularity and edit volume are well-taken, and we address each major comment below with plans for revision.

read point-by-point responses

Referee: [Abstract and temporal analysis] Abstract and temporal analysis section: the reported shorter IETs for transitions between hyperlinked article pairs versus non-hyperlinked pairs do not include controls for article popularity or edit frequency. Popular articles receive more edits and are more likely to be hyperlinked, so the IET difference could arise from a popularity-driven null rather than hyperlink structure itself. A degree-preserving rewiring null model or regression adjustment for edit volume is needed to support the structural claim.

Authors: We agree that article popularity and edit frequency represent plausible confounds that could contribute to the observed differences in inter-event times. To isolate the contribution of hyperlink structure, we will incorporate a degree-preserving rewiring null model in the revised manuscript. This null model will rewire the hyperlink network while preserving the degree sequence, allowing us to compare empirical IET distributions against those expected under degree-matched randomization. We will report the results of this test, including statistical significance of the IET reduction for hyperlinked pairs relative to the null, in a new subsection of the temporal analysis. revision: yes
Referee: [Jaccard similarity analysis] Jaccard similarity analysis: the higher Jaccard overlap between specialists' transition networks and hyperlink subnetworks may likewise be driven by the correlation between hyperlink degree and edit volume. Without explicit degree-matching or null models that rewire hyperlinks while preserving the degree sequence, the measure does not isolate the influence of the underlying link structure from confounding editor-article popularity effects.

Authors: We concur that the Jaccard similarity metric could be inflated by correlations between hyperlink degree and edit volume. In the revision, we will add robustness checks that apply degree-preserving rewiring to the hyperlink network and generate null distributions for the Jaccard values. We will demonstrate that the elevated Jaccard similarity observed for specialist editors remains statistically distinguishable from the null expectation. These additional controls will be integrated into the editor classification section to strengthen the claim that the alignment reflects hyperlink structure rather than popularity effects alone. revision: yes

Circularity Check

0 steps flagged

No circularity: all measures computed directly from raw edit logs and hyperlink data

full rationale

The paper performs purely observational comparisons: inter-event times are measured between consecutive edits in the actual editing sequences, Jaccard similarity is calculated between each editor's observed transition graph and the hyperlink graph on the same articles, and communities are detected via standard algorithms on the hyperlink network. No parameters are fitted to a subset of the data and then re-used as a 'prediction'; no equations define one quantity in terms of another that is then claimed as an independent result; and no load-bearing steps rely on self-citations whose content is itself unverified or defined by the present work. The analysis remains self-contained and externally falsifiable against the Wikipedia edit history and link structure.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard assumptions about hyperlinks as meaningful connections and the validity of community detection plus similarity metrics; no new entities are postulated and the 19 communities represent a modeling choice.

free parameters (1)

Number of topical communities = 19
Coarse-graining the hyperlink network into exactly 19 topical communities; chosen to organize the structural view of editing behavior.

axioms (1)

domain assumption Hyperlinks between articles represent connections that can influence the sequence of editorial edits.
Invoked when interpreting shorter IETs for hyperlinked pairs as evidence of structural closeness in editing.

pith-pipeline@v0.9.0 · 5835 in / 1298 out tokens · 55286 ms · 2026-05-19T19:34:40.426937+00:00 · methodology

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

transitions between hyperlinked article pairs have shorter inter-event times (IETs) than those between non-hyperlinked pairs... Jaccard similarity

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

27 extracted references · 27 canonical work pages

[1]

Measuring Wikipedia

V oss, J. Measuring Wikipedia. InProc. 10th Int. Conf. Scientom. Informetr., vol. 1, 221–231 (Stockholm, Sweden, 2005)

work page 2005
[2]

& Kertész, J

Yasseri, T., Sumi, R., Rung, A., Kornai, A. & Kertész, J. Dynamics of conflicts in Wikipedia.PLOS ONE7,e38869, DOI: 10.1371/journal.pone.0038869 (2012)

work page doi:10.1371/journal.pone.0038869 2012
[3]

& Shimada, T

Ogushi, F., Kertész, J., Kaski, K. & Shimada, T. Ecology of the digital world of Wikipedia.Sci. Rep.11,18371, DOI: 10.1038/s41598-021-97755-w (2021). 4.Giles, J. Internet encyclopaedias go head to head.Nature438,900–901, DOI: 10.1038/438900a (2005)

work page doi:10.1038/s41598-021-97755-w 2021
[4]

Son, S.-W.et al.Sampling properties of directed networks.Phys. Rev. E86,046104, DOI: 10.1103/PhysRevE.86.046104 (2012)

work page doi:10.1103/physreve.86.046104 2012
[5]

E., Saam, M., Halatchliyski, I

Kummer, M. E., Saam, M., Halatchliyski, I. & Giorgidze, G. Centrality and content creation in networks - The case of economic topics on German Wikipedia.Inf. Econ. Policy36,36–52, DOI: 10.1016/j.infoecopol.2016.06.002 (2016)

work page doi:10.1016/j.infoecopol.2016.06.002 2016
[6]

Schwartz, G. A. Complex networks reveal emergent interdisciplinary knowledge in Wikipedia.Humanit. Soc. Sci. Commun. 8,127, DOI: 10.1057/s41599-021-00801-1 (2021)

work page doi:10.1057/s41599-021-00801-1 2021
[7]

& Strohmaier, M

Lamprecht, D., Lerman, K., Helic, D. & Strohmaier, M. How the structure of Wikipedia articles influences user navigation. New Rev. Hypermedia Multimed.23,29–50, DOI: 10.1080/13614568.2016.1179798 (2017)

work page doi:10.1080/13614568.2016.1179798 2016
[8]

& Leskovec, J

West, R. & Leskovec, J. Human wayfinding in information networks. InProc. 21st Int. Conf. World Wide Web (WWW ’12), 619–628, DOI: 10.1145/2187836.2187920 (Association for Computing Machinery, New York, NY , USA, 2012)

work page doi:10.1145/2187836.2187920 2012
[9]

Analyzing user click paths in a Wikipedia navigation game

Helic, D. Analyzing user click paths in a Wikipedia navigation game. InProc. 35th Int. Conv. MIPRO, 374–379, (IEEE, 2012)

work page 2012
[10]

& Leskovec, J

West, R., Paranjape, A. & Leskovec, J. Mining missing hyperlinks from human navigation traces: A case study of Wikipedia. InProc. 24th Int. Conf. World Wide Web (WWW ’15), 1242–1252, DOI: 10.1145/2736277.2741666 (Association for Computing Machinery, New York, NY , USA, 2015)

work page doi:10.1145/2736277.2741666 2015
[11]

& Kertész, J

Zhu, M. & Kertész, J. Milgram’s experiment in the knowledge space: individual navigation strategies.EPJ Data Sci.14, 42, DOI: 10.1140/epjds/s13688-025-00558-6 (2025)

work page doi:10.1140/epjds/s13688-025-00558-6 2025
[12]

Singer, P.et al.Why we read Wikipedia. InProc. 26th Int. Conf. World Wide Web (WWW ’17), 1591–1600, DOI: 10.1145/3038912.3052716 (International World Wide Web Conferences Steering Committee, Republic and Canton of Geneva, CHE, 2017)

work page doi:10.1145/3038912.3052716 2017
[13]

Wei, B.et al.Motif-based hyponym relation extraction from Wikipedia hyperlinks.IEEE Trans. Knowl. Data Eng.26, 2507–2519, DOI: 10.1109/TKDE.2013.183 (2014)

work page doi:10.1109/tkde.2013.183 2013
[14]

& Lee, C

Kim, J., Kim, S. & Lee, C. Anticipating technological convergence: Link prediction using Wikipedia hyperlinks. Technovation79,25–34, DOI: 10.1016/j.technovation.2018.06.008 (2019)

work page doi:10.1016/j.technovation.2018.06.008 2018
[15]

Wei, B.et al.DF-Miner: Domain-specific facet mining by leveraging the hyperlink structure of Wikipedia.Knowl.-Based Syst.77,80–91, DOI: 10.1016/j.knosys.2015.01.001 (2015)

work page doi:10.1016/j.knosys.2015.01.001 2015
[16]

Witten, I. H. & Milne, D. N. An effective, low-cost measure of semantic relatedness obtained from Wikipedia links. In Proc. AAAI Workshop on Wikipedia and Artificial Intelligence: An Evolving Synergy, 25–30 (AAAI Press, Chicago, IL, USA, 2008)

work page 2008
[17]

Geiger, R. S. & Halfaker, A. Using edit sessions to measure participation in Wikipedia. InProc. 2013 Conf. Comput. Support. Coop. Work (CSCW ’13), 861–870, DOI: 10.1145/2441776.2441873 (ACM, 2013)

work page doi:10.1145/2441776.2441873 2013
[18]

Yun, J., Lee, S. H. & Jeong, H. Early onset of structural inequality in the formation of collaborative knowledge in all Wikimedia projects.Nat. Hum. Behav.3,155–163, DOI: 10.1038/s41562-018-0488-z (2019)

work page doi:10.1038/s41562-018-0488-z 2019
[19]

& Shimada, T

Ogushi, F. & Shimada, T. Comparison of metrics for measuring Wikipedia ecology: characteristics of self-consistent metrics for editor scatteredness and article complexity.Artif. Life Robotics28,62–66, DOI: 10.1007/s10015-022-00819-x (2023). 12/18

work page doi:10.1007/s10015-022-00819-x 2023
[20]

& Jung, W.-S

Kwon, O., Son, W.-S. & Jung, W.-S. The double power law in human collaboration behavior: The case of Wikipedia.Phys. A: Stat. Mech. Appl.461,85–91, DOI: 10.1016/j.physa.2016.05.010 (2016)

work page doi:10.1016/j.physa.2016.05.010 2016
[21]

& Jo, H.-H

Choi, J., Hiraoka, T. & Jo, H.-H. Individual-driven versus interaction-driven burstiness in human dynamics: The case of Wikipedia edit history.Phys. Rev. E104,014312, DOI: 10.1103/PhysRevE.104.014312 (2021)

work page doi:10.1103/physreve.104.014312 2021
[22]

& Kaski, K

Shimada, T., Ogushi, F., Török, J., Kertész, J. & Kaski, K. A simple model of edit activity in Wikipedia.Phys. A: Stat. Mech. Appl.630,129253, DOI: 10.1016/j.physa.2023.129253 (2023). 24.Wikipedia. Category: All Wikipedia bots. https://en.wikipedia.org/wiki/Category:All_Wikipedia_bots (2025). Accessed 28 Oct 2025

work page doi:10.1016/j.physa.2023.129253 2023
[23]

Yun, J., Lee, S. H. & Jeong, H. Intellectual interchanges in the history of the massive online open-editing encyclopedia, Wikipedia.Phys. Rev. E93,012307, DOI: 10.1103/PhysRevE.93.012307 (2016)

work page doi:10.1103/physreve.93.012307 2016
[24]

Traag, Ludo Waltman, and Nees Jan van Eck

Traag, V .A., Waltman, L. & van Eck, N.J. From Louvain to Leiden: guaranteeing well-connected communities.Sci. Rep.9, 5233, DOI: 10.1038/s41598-019-41695-z (2019). 27.Wikimedia Foundation. Wikimedia Downloads. https://dumps.wikimedia.org/ (2025). Accessed 24 Apr 2025

work page doi:10.1038/s41598-019-41695-z 2019
[25]

The origin of bursts and heavy tails in human dynamics.Nature435,207–211, DOI: 10.1038/nature03459 (2005)

Barabási, A.-L. The origin of bursts and heavy tails in human dynamics.Nature435,207–211, DOI: 10.1038/nature03459 (2005)

work page doi:10.1038/nature03459 2005
[26]

& Kaski, K.Bursty Human DynamicsDOI: 10.1007/978-3-319-68540-3 (Springer, Cham, Switzer- land, 2018)

Karsai, M., Jo, H.-H. & Kaski, K.Bursty Human DynamicsDOI: 10.1007/978-3-319-68540-3 (Springer, Cham, Switzer- land, 2018). 30.Simpson, E. Measurement of Diversity.Nature163,688, DOI: 10.1038/163688a0 (1949)

work page doi:10.1007/978-3-319-68540-3 2018
[27]

Planetary Resilience Science for Safeguarding the Global Commons

Karsai, M., Kaski, K., Barabási, A.-L. & Kertész, J. Universal features of correlated bursty behaviour.Sci. Rep.2,397, DOI: 10.1038/srep00397 (2012). Acknowledgements We thank Dr. Jong-Min Park for the fruitful discussion. This work was supported by the National Research Foundation (NRF) of Korea through Grant Numbers. RS-2024-00341317 (M.J.L.), RS-2026-2...

work page doi:10.1038/srep00397 2012

[1] [1]

Measuring Wikipedia

V oss, J. Measuring Wikipedia. InProc. 10th Int. Conf. Scientom. Informetr., vol. 1, 221–231 (Stockholm, Sweden, 2005)

work page 2005

[2] [2]

& Kertész, J

Yasseri, T., Sumi, R., Rung, A., Kornai, A. & Kertész, J. Dynamics of conflicts in Wikipedia.PLOS ONE7,e38869, DOI: 10.1371/journal.pone.0038869 (2012)

work page doi:10.1371/journal.pone.0038869 2012

[3] [3]

& Shimada, T

Ogushi, F., Kertész, J., Kaski, K. & Shimada, T. Ecology of the digital world of Wikipedia.Sci. Rep.11,18371, DOI: 10.1038/s41598-021-97755-w (2021). 4.Giles, J. Internet encyclopaedias go head to head.Nature438,900–901, DOI: 10.1038/438900a (2005)

work page doi:10.1038/s41598-021-97755-w 2021

[4] [4]

Son, S.-W.et al.Sampling properties of directed networks.Phys. Rev. E86,046104, DOI: 10.1103/PhysRevE.86.046104 (2012)

work page doi:10.1103/physreve.86.046104 2012

[5] [5]

E., Saam, M., Halatchliyski, I

Kummer, M. E., Saam, M., Halatchliyski, I. & Giorgidze, G. Centrality and content creation in networks - The case of economic topics on German Wikipedia.Inf. Econ. Policy36,36–52, DOI: 10.1016/j.infoecopol.2016.06.002 (2016)

work page doi:10.1016/j.infoecopol.2016.06.002 2016

[6] [6]

Schwartz, G. A. Complex networks reveal emergent interdisciplinary knowledge in Wikipedia.Humanit. Soc. Sci. Commun. 8,127, DOI: 10.1057/s41599-021-00801-1 (2021)

work page doi:10.1057/s41599-021-00801-1 2021

[7] [7]

& Strohmaier, M

Lamprecht, D., Lerman, K., Helic, D. & Strohmaier, M. How the structure of Wikipedia articles influences user navigation. New Rev. Hypermedia Multimed.23,29–50, DOI: 10.1080/13614568.2016.1179798 (2017)

work page doi:10.1080/13614568.2016.1179798 2016

[8] [8]

& Leskovec, J

West, R. & Leskovec, J. Human wayfinding in information networks. InProc. 21st Int. Conf. World Wide Web (WWW ’12), 619–628, DOI: 10.1145/2187836.2187920 (Association for Computing Machinery, New York, NY , USA, 2012)

work page doi:10.1145/2187836.2187920 2012

[9] [9]

Analyzing user click paths in a Wikipedia navigation game

Helic, D. Analyzing user click paths in a Wikipedia navigation game. InProc. 35th Int. Conv. MIPRO, 374–379, (IEEE, 2012)

work page 2012

[10] [10]

& Leskovec, J

West, R., Paranjape, A. & Leskovec, J. Mining missing hyperlinks from human navigation traces: A case study of Wikipedia. InProc. 24th Int. Conf. World Wide Web (WWW ’15), 1242–1252, DOI: 10.1145/2736277.2741666 (Association for Computing Machinery, New York, NY , USA, 2015)

work page doi:10.1145/2736277.2741666 2015

[11] [11]

& Kertész, J

Zhu, M. & Kertész, J. Milgram’s experiment in the knowledge space: individual navigation strategies.EPJ Data Sci.14, 42, DOI: 10.1140/epjds/s13688-025-00558-6 (2025)

work page doi:10.1140/epjds/s13688-025-00558-6 2025

[12] [12]

Singer, P.et al.Why we read Wikipedia. InProc. 26th Int. Conf. World Wide Web (WWW ’17), 1591–1600, DOI: 10.1145/3038912.3052716 (International World Wide Web Conferences Steering Committee, Republic and Canton of Geneva, CHE, 2017)

work page doi:10.1145/3038912.3052716 2017

[13] [13]

Wei, B.et al.Motif-based hyponym relation extraction from Wikipedia hyperlinks.IEEE Trans. Knowl. Data Eng.26, 2507–2519, DOI: 10.1109/TKDE.2013.183 (2014)

work page doi:10.1109/tkde.2013.183 2013

[14] [14]

& Lee, C

Kim, J., Kim, S. & Lee, C. Anticipating technological convergence: Link prediction using Wikipedia hyperlinks. Technovation79,25–34, DOI: 10.1016/j.technovation.2018.06.008 (2019)

work page doi:10.1016/j.technovation.2018.06.008 2018

[15] [15]

Wei, B.et al.DF-Miner: Domain-specific facet mining by leveraging the hyperlink structure of Wikipedia.Knowl.-Based Syst.77,80–91, DOI: 10.1016/j.knosys.2015.01.001 (2015)

work page doi:10.1016/j.knosys.2015.01.001 2015

[16] [16]

Witten, I. H. & Milne, D. N. An effective, low-cost measure of semantic relatedness obtained from Wikipedia links. In Proc. AAAI Workshop on Wikipedia and Artificial Intelligence: An Evolving Synergy, 25–30 (AAAI Press, Chicago, IL, USA, 2008)

work page 2008

[17] [17]

Geiger, R. S. & Halfaker, A. Using edit sessions to measure participation in Wikipedia. InProc. 2013 Conf. Comput. Support. Coop. Work (CSCW ’13), 861–870, DOI: 10.1145/2441776.2441873 (ACM, 2013)

work page doi:10.1145/2441776.2441873 2013

[18] [18]

Yun, J., Lee, S. H. & Jeong, H. Early onset of structural inequality in the formation of collaborative knowledge in all Wikimedia projects.Nat. Hum. Behav.3,155–163, DOI: 10.1038/s41562-018-0488-z (2019)

work page doi:10.1038/s41562-018-0488-z 2019

[19] [19]

& Shimada, T

Ogushi, F. & Shimada, T. Comparison of metrics for measuring Wikipedia ecology: characteristics of self-consistent metrics for editor scatteredness and article complexity.Artif. Life Robotics28,62–66, DOI: 10.1007/s10015-022-00819-x (2023). 12/18

work page doi:10.1007/s10015-022-00819-x 2023

[20] [20]

& Jung, W.-S

Kwon, O., Son, W.-S. & Jung, W.-S. The double power law in human collaboration behavior: The case of Wikipedia.Phys. A: Stat. Mech. Appl.461,85–91, DOI: 10.1016/j.physa.2016.05.010 (2016)

work page doi:10.1016/j.physa.2016.05.010 2016

[21] [21]

& Jo, H.-H

Choi, J., Hiraoka, T. & Jo, H.-H. Individual-driven versus interaction-driven burstiness in human dynamics: The case of Wikipedia edit history.Phys. Rev. E104,014312, DOI: 10.1103/PhysRevE.104.014312 (2021)

work page doi:10.1103/physreve.104.014312 2021

[22] [22]

& Kaski, K

Shimada, T., Ogushi, F., Török, J., Kertész, J. & Kaski, K. A simple model of edit activity in Wikipedia.Phys. A: Stat. Mech. Appl.630,129253, DOI: 10.1016/j.physa.2023.129253 (2023). 24.Wikipedia. Category: All Wikipedia bots. https://en.wikipedia.org/wiki/Category:All_Wikipedia_bots (2025). Accessed 28 Oct 2025

work page doi:10.1016/j.physa.2023.129253 2023

[23] [23]

Yun, J., Lee, S. H. & Jeong, H. Intellectual interchanges in the history of the massive online open-editing encyclopedia, Wikipedia.Phys. Rev. E93,012307, DOI: 10.1103/PhysRevE.93.012307 (2016)

work page doi:10.1103/physreve.93.012307 2016

[24] [24]

Traag, Ludo Waltman, and Nees Jan van Eck

Traag, V .A., Waltman, L. & van Eck, N.J. From Louvain to Leiden: guaranteeing well-connected communities.Sci. Rep.9, 5233, DOI: 10.1038/s41598-019-41695-z (2019). 27.Wikimedia Foundation. Wikimedia Downloads. https://dumps.wikimedia.org/ (2025). Accessed 24 Apr 2025

work page doi:10.1038/s41598-019-41695-z 2019

[25] [25]

The origin of bursts and heavy tails in human dynamics.Nature435,207–211, DOI: 10.1038/nature03459 (2005)

Barabási, A.-L. The origin of bursts and heavy tails in human dynamics.Nature435,207–211, DOI: 10.1038/nature03459 (2005)

work page doi:10.1038/nature03459 2005

[26] [26]

& Kaski, K.Bursty Human DynamicsDOI: 10.1007/978-3-319-68540-3 (Springer, Cham, Switzer- land, 2018)

Karsai, M., Jo, H.-H. & Kaski, K.Bursty Human DynamicsDOI: 10.1007/978-3-319-68540-3 (Springer, Cham, Switzer- land, 2018). 30.Simpson, E. Measurement of Diversity.Nature163,688, DOI: 10.1038/163688a0 (1949)

work page doi:10.1007/978-3-319-68540-3 2018

[27] [27]

Planetary Resilience Science for Safeguarding the Global Commons

Karsai, M., Kaski, K., Barabási, A.-L. & Kertész, J. Universal features of correlated bursty behaviour.Sci. Rep.2,397, DOI: 10.1038/srep00397 (2012). Acknowledgements We thank Dr. Jong-Min Park for the fruitful discussion. This work was supported by the National Research Foundation (NRF) of Korea through Grant Numbers. RS-2024-00341317 (M.J.L.), RS-2026-2...

work page doi:10.1038/srep00397 2012