pith. sign in

arxiv: 2604.04538 · v1 · submitted 2026-04-06 · ⚛️ physics.soc-ph

Disentangling Large-Scale Supply Networks: f-HiCoNE Framework for Flow-Hierarchical Clustering via Combinatorial Hodge Decomposition

Taiyo Nakatani , Takaaki Aoki This is my paper

Pith reviewed 2026-05-10 19:31 UTC · model grok-4.3

classification ⚛️ physics.soc-ph
keywords supply chain networksflow-hierarchical clusteringcombinatorial Hodge decompositioninter-firm transactionsacyclic gradient flowindustrial ecosystemsnetwork partitioning
0
0 comments X

The pith

Combinatorial Hodge decomposition isolates acyclic flows in transaction networks to extract flow-hierarchical supply chain clusters.

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

The paper introduces the f-HiCoNE framework to handle the fact that real inter-firm transaction networks contain many cycles even though supply chains are conceptually hierarchical. Combinatorial Hodge decomposition separates the network flow into an acyclic gradient component that quantifies upstream-downstream order. This component is then used to partition the graph into clusters. When applied to a dataset covering roughly 650,000 firms, the resulting clusters show clear hierarchical organisation captured by local scalar potentials and correspond to geographically localised industrial ecosystems. A reader would care because the method supplies a concrete way to recover latent hierarchy from otherwise entangled large-scale transaction data.

Core claim

By applying combinatorial Hodge decomposition to inter-firm transaction graphs, the f-HiCoNE framework isolates the acyclic gradient flow, quantifies the flow-hierarchical parts, and partitions the graph into functional supply-chain clusters. In a nationwide dataset of approximately 650,000 firms these clusters exhibit strong flow-hierarchical organisation, with upstream-downstream positioning of firms accurately captured by local scalar potentials, revealing distinct geographically localised industrial ecosystems.

What carries the argument

combinatorial Hodge decomposition on the transaction graph, which isolates the acyclic gradient flow component used to quantify hierarchy and perform the partitioning.

If this is right

  • The framework partitions massive cyclic transaction networks into clusters that preserve upstream-downstream order.
  • Scalar potentials derived from the gradient flow give a quantitative measure of each firm's position in the hierarchy.
  • The resulting clusters correspond to geographically localised industrial ecosystems.
  • The approach supplies firms with a map of their position within the larger inter-firm network.

Where Pith is reading between the lines

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

  • The same decomposition could be applied to other large directed networks where an underlying acyclic structure is suspected beneath cycles.
  • If the scalar potentials prove stable under removal of individual edges, the hierarchy may reflect robust economic dependencies rather than data noise.
  • Extending the method to time-varying transaction data could reveal how hierarchical clusters evolve during economic shocks.

Load-bearing premise

That the acyclic gradient flow isolated by the decomposition corresponds to meaningful real-world supply-chain hierarchy rather than an artifact of preprocessing or the decomposition itself.

What would settle it

Check whether the extracted local scalar potentials correlate with independent indicators of firm position such as known production stages or supplier-customer links in a held-out validation subset of firms.

Figures

Figures reproduced from arXiv: 2604.04538 by Taiyo Nakatani, Takaaki Aoki.

Figure 1
Figure 1. Figure 1: (A) The concept of supply chain emphasizes a hierarchical, ordered flow from upstream suppliers through intermediate stages to downstream assemblers. (B) Real-world inter-firm data rarely resembles an acyclic, flow-hierarchical structure. (C) We propose flow-Hierarchical clustering to recover the hierarchical tiers and upstream–downstream positioning of individual firms within the clusters. However, supply… view at source ↗
Figure 2
Figure 2. Figure 2: Combinatorial Hodge decomposition isolates the acyclic component from a given inter-firm [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: The flow-Hierarchical Community Network Extraction (f-HiCoNE) algorithm. The process consists of three stages: (Step 1) Quantifying the connection strength of firms with the acyclic flow to define new edge weights wij ; (Step 2) Partitioning the graph G(V, E, W) into clusters dominated by the acyclic flow using these weights; and (Step 3) Calculating the scalar potential within each extracted cluster to de… view at source ↗
Figure 4
Figure 4. Figure 4: Characteristics of extracted flow-hierarchical clusters. (A) Cluster sizes (number of firms). (B) Distribution of the coefficient of determination, R2 . The red vertical line indicates the R2 of the entire network. gradient component obtained this decomposition. By replacing the original edge flows Y with W, we obtain an undirected, weighted graph G′(V, E, W). Step 2 extracts flow-hierarchical clusters fro… view at source ↗
Figure 5
Figure 5. Figure 5: Industrial and geographical characteristics of Cluster 1. (A) Industrial composition visualised as a treemap, grouped by JSIC 2-digit sectors. Rectangle area is proportional to the sector’s total transaction amount within the cluster, and colour represents the median of the scalar potentials (red: upstream/high potential; blue: downstream/low potential). (B) Relative geographical density of firm headquarte… view at source ↗
Figure 6
Figure 6. Figure 6: Industrial and geographic characteristics of Cluster 2. Panels (A)–(B) and the visual encodings are defined as in [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Industrial and geographic characteristics of Cluster 3. Panels (A)–(B) and the visual encodings are defined as in [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Industrial and geographic characteristics of Cluster 4. Panels [PITH_FULL_IMAGE:figures/full_fig_p014_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Industrial and geographic characteristics of Cluster 5. Panels [PITH_FULL_IMAGE:figures/full_fig_p015_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Industrial and geographic characteristics of Cluster 6. Panels [PITH_FULL_IMAGE:figures/full_fig_p015_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Industrial and geographic characteristics of Cluster 7. Panels [PITH_FULL_IMAGE:figures/full_fig_p016_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Industrial and geographic characteristics of Cluster 8. Panels [PITH_FULL_IMAGE:figures/full_fig_p017_12.png] view at source ↗
Figure 13
Figure 13. Figure 13: Industrial and geographic characteristics of Cluster 9. Panels [PITH_FULL_IMAGE:figures/full_fig_p017_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Industrial and geographic characteristics of Cluster 10. Panels [PITH_FULL_IMAGE:figures/full_fig_p018_14.png] view at source ↗
Figure 15
Figure 15. Figure 15: Industrial and geographic characteristics of Cluster 11. Panels [PITH_FULL_IMAGE:figures/full_fig_p018_15.png] view at source ↗
Figure 16
Figure 16. Figure 16: Industrial and geographic characteristics of Cluster 12. Panels [PITH_FULL_IMAGE:figures/full_fig_p019_16.png] view at source ↗
Figure 17
Figure 17. Figure 17: Industrial and geographic characteristics of Cluster 13. Panels [PITH_FULL_IMAGE:figures/full_fig_p020_17.png] view at source ↗
Figure 18
Figure 18. Figure 18: Industrial and geographic characteristics of Cluster 14. Panels [PITH_FULL_IMAGE:figures/full_fig_p020_18.png] view at source ↗
Figure 19
Figure 19. Figure 19: Industrial and geographic characteristics of Cluster 15. Panels [PITH_FULL_IMAGE:figures/full_fig_p021_19.png] view at source ↗
Figure 20
Figure 20. Figure 20: Industrial and geographic characteristics of Cluster 16. Panels [PITH_FULL_IMAGE:figures/full_fig_p022_20.png] view at source ↗
Figure 21
Figure 21. Figure 21: Industrial and geographic characteristics of Cluster 17. Panels [PITH_FULL_IMAGE:figures/full_fig_p022_21.png] view at source ↗
Figure 22
Figure 22. Figure 22: Industrial and geographic characteristics of Cluster 18. Panels [PITH_FULL_IMAGE:figures/full_fig_p023_22.png] view at source ↗
Figure 23
Figure 23. Figure 23: Industrial and geographic characteristics of Cluster 19. Panels [PITH_FULL_IMAGE:figures/full_fig_p023_23.png] view at source ↗
Figure 24
Figure 24. Figure 24: Industrial and geographic characteristics of Cluster 20. Panels [PITH_FULL_IMAGE:figures/full_fig_p024_24.png] view at source ↗
Figure 25
Figure 25. Figure 25: Industrial and geographic characteristics of Cluster 21. Panels [PITH_FULL_IMAGE:figures/full_fig_p024_25.png] view at source ↗
Figure 26
Figure 26. Figure 26: Industrial and geographic characteristics of Cluster 22. Panels [PITH_FULL_IMAGE:figures/full_fig_p025_26.png] view at source ↗
Figure 27
Figure 27. Figure 27: Industrial and geographic characteristics of Cluster 23. Panels [PITH_FULL_IMAGE:figures/full_fig_p025_27.png] view at source ↗
Figure 28
Figure 28. Figure 28: Industrial and geographic characteristics of Cluster 24. Panels [PITH_FULL_IMAGE:figures/full_fig_p026_28.png] view at source ↗
Figure 29
Figure 29. Figure 29: Industrial and geographic characteristics of Cluster 25. Panels [PITH_FULL_IMAGE:figures/full_fig_p026_29.png] view at source ↗
Figure 30
Figure 30. Figure 30: Industrial and geographic characteristics of Cluster 26. Panels [PITH_FULL_IMAGE:figures/full_fig_p027_30.png] view at source ↗
Figure 31
Figure 31. Figure 31: Industrial and geographic characteristics of Cluster 27. Panels [PITH_FULL_IMAGE:figures/full_fig_p027_31.png] view at source ↗
read the original abstract

Modern society relies on complex supply chains to sustain the flow of goods and services that are essential to daily life. While traditional supply chain theory assumes a clear, hierarchical flow from upstream suppliers to downstream customers, observable real-world transaction networks rarely exhibit this acyclic structure. Instead, detailed inter-firm data reveal that interwoven networks are heavily entangled by cyclic flows. Consequently, without appropriate partitioning of these massive inter-firm networks, the latent flow-hierarchical structures that are central to supply chain concepts remain obscure. To address this analytical challenge, we introduce the flow-Hierarchical Community Network Extraction (f-HiCoNE) framework. By applying combinatorial Hodge decomposition, this approach disentangles the complex inter-firm network by isolating the acyclic gradient flow to quantify the flow-hierarchical parts and partition the graph. By applying f-HiCoNE to a nationwide transaction dataset of approximately 650,000 firms, we successfully extracted functional supply-chain clusters. These clusters demonstrated strong flow-hierarchical organisation, wherein the upstream-downstream positioning of firms was accurately captured by local scalar potentials, revealing distinct geographically localised industrial ecosystems. This study provides a map that helps firms understand their surrounding environment and locate their position within an inter-firm network and opens a new research avenue focused on flow-hierarchy clustering in supply chain analysis.

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.

Referee Report

3 major / 0 minor

Summary. The paper introduces the f-HiCoNE framework, which applies combinatorial Hodge decomposition to large directed transaction graphs to isolate the acyclic gradient flow component. This yields a scalar potential used to quantify flow hierarchy and extract supply-chain clusters. Applied to a nationwide dataset of ~650,000 firms, the authors claim the resulting clusters exhibit strong flow-hierarchical organization, with local potentials accurately capturing upstream-downstream firm positioning and revealing geographically localized industrial ecosystems.

Significance. If externally validated, the framework would supply a mathematically principled method for disentangling cyclic flows in economic networks and recovering latent hierarchy, extending combinatorial Hodge theory to supply-chain analysis and offering practical mapping tools for firms. The scale of the real-world application is a positive feature, but the current lack of quantitative validation metrics or baseline comparisons substantially limits demonstrated significance.

major comments (3)
  1. Abstract: the central claim that 'the upstream-downstream positioning of firms was accurately captured by local scalar potentials' is presented without any validation metrics, baseline comparisons, error analysis, or description of how post-decomposition clustering decisions were made. This absence directly undermines assessment of the 'successful extraction' and 'strong flow-hierarchical organisation' assertions.
  2. The gradient component is acyclic by construction of the orthogonal decomposition; the manuscript therefore needs to show that the resulting potentials align with independent economic indicators of hierarchy (e.g., known supplier-customer relations or industry classifications) rather than reflecting an imposed feature of the projection or preprocessing. No such external alignment test is described.
  3. The weakest assumption—that the isolated acyclic gradient flow corresponds to meaningful real-world supply-chain hierarchy—is load-bearing for the interpretation of the clusters as 'functional' and 'geographically localised industrial ecosystems,' yet remains untested against alternative explanations such as data encoding choices (gross vs. net volumes).

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed comments, which highlight important aspects of validation and interpretation. We agree that additional quantitative support and explicit discussion of assumptions will strengthen the manuscript. We will incorporate revisions to address these points while preserving the core contribution of applying combinatorial Hodge decomposition to disentangle supply networks. Our point-by-point responses follow.

read point-by-point responses

  1. Referee: Abstract: the central claim that 'the upstream-downstream positioning of firms was accurately captured by local scalar potentials' is presented without any validation metrics, baseline comparisons, error analysis, or description of how post-decomposition clustering decisions were made. This absence directly undermines assessment of the 'successful extraction' and 'strong flow-hierarchical organisation' assertions.

    Authors: We acknowledge that the abstract is highly condensed and does not reference the supporting analyses. The full manuscript presents the alignment through visualizations of scalar potentials overlaid on transaction flows, cluster coherence based on intra-cluster flow directionality, and geographic consistency checks. To address the concern directly, we will revise the abstract to include a brief statement on the validation approach and clustering procedure (e.g., potential thresholding and modularity on the gradient component). We will also expand the methods section with a short description of how post-decomposition decisions were made. revision: yes

  2. Referee: The gradient component is acyclic by construction of the orthogonal decomposition; the manuscript therefore needs to show that the resulting potentials align with independent economic indicators of hierarchy (e.g., known supplier-customer relations or industry classifications) rather than reflecting an imposed feature of the projection or preprocessing. No such external alignment test is described.

    Authors: We agree that alignment with external indicators is necessary to establish economic meaning beyond the mathematical construction. The manuscript already shows that the extracted potentials produce geographically localized clusters consistent with known industrial districts and that potential values correlate with firm attributes such as sector and transaction volume. However, we did not report explicit quantitative correlations with independent supplier lists. In revision we will add a dedicated subsection presenting Spearman correlations between local potentials and industry classifications plus a comparison against a simple baseline (e.g., degree-based ranking), thereby providing the requested external validation. revision: yes

  3. Referee: The weakest assumption—that the isolated acyclic gradient flow corresponds to meaningful real-world supply-chain hierarchy—is load-bearing for the interpretation of the clusters as 'functional' and 'geographically localised industrial ecosystems,' yet remains untested against alternative explanations such as data encoding choices (gross vs. net volumes).

    Authors: This is a fair observation on the interpretive step. The framework isolates the gradient component precisely to recover the hierarchical signal that traditional cycle-heavy graphs obscure; the resulting clusters' strong geographic localization and flow-direction consistency provide supporting evidence for the supply-chain interpretation. We performed robustness checks on edge thresholding but did not explicitly contrast gross versus net transaction volumes. We will add a limitations paragraph discussing this assumption, including why the orthogonal decomposition mitigates preprocessing artifacts, and note that gross/net sensitivity can be explored in future work with alternative datasets. We therefore revise partially by expanding the discussion rather than performing an entirely new experiment. revision: partial

Circularity Check

1 steps flagged

Scalar potentials from Hodge decomposition define flow hierarchy by construction

specific steps
  1. self definitional [Abstract]
    "These clusters demonstrated strong flow-hierarchical organisation, wherein the upstream-downstream positioning of firms was accurately captured by local scalar potentials, revealing distinct geographically localised industrial ecosystems."

    The local scalar potentials are obtained directly from the gradient component of the combinatorial Hodge decomposition applied to the same transaction graph. By the properties of the decomposition, this component is always acyclic and equals the discrete gradient of a scalar function; therefore the claim that the potentials 'accurately capture' upstream-downstream positioning is equivalent to the mathematical definition of the extracted flow rather than a separate validation.

full rationale

The paper's central result applies combinatorial Hodge decomposition to isolate an acyclic gradient flow on the transaction graph and interprets the resulting node potentials as capturing real upstream-downstream supply-chain positioning. Because the gradient component is mathematically guaranteed to be acyclic and exactly representable as potential differences (by the definition of the orthogonal decomposition), the reported 'accurate capture' of hierarchy reduces directly to the input projection rather than an independent empirical finding. The abstract provides the load-bearing claim but no external alignment check against known supply-chain data is quoted. This produces moderate circularity confined to the interpretation step; the clustering algorithm itself remains a non-circular application of the decomposition.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The central claim rests on the applicability of combinatorial Hodge decomposition to directed transaction graphs and on the interpretation of the resulting gradient potentials as upstream-downstream positions; no explicit free parameters, ad-hoc axioms, or new entities are stated in the abstract.

pith-pipeline@v0.9.0 · 5544 in / 1141 out tokens · 59599 ms · 2026-05-10T19:31:37.793999+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

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

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

35 extracted references · 35 canonical work pages

  1. [1]

    & Choi, T

    Mena, C., Humphries, A. & Choi, T. Y. Toward a Theory of Multi-Tier Supply Chain Management. Journal of Supply Chain Management49,58–77. doi:10.1111/jscm.12003(2013)

    work page doi:10.1111/jscm.12003(2013 2013

  2. [2]

    Bellamy, M. A. & Basole, R. C. Network Analysis of Supply Chain Systems: A Systematic Review and Future Research.Systems Engineering16,235–249 (2013)

    work page 2013

  3. [3]

    & Thurner, S

    Diem, C., Borsos, A., Reisch, T., Kert´ esz, J. & Thurner, S. Estimating the Loss of Economic Predictability from Aggregating Firm-Level Production Networks.PNAS Nexus3,pgae064. doi:10. 1093/pnasnexus/pgae064(Mar. 2024)

    work page 2024

  4. [4]

    Borgatti, S. P. & Li, X. On Social Network Analysis in a Supply Chain Context.Journal of Supply Chain Management45,5–22. doi:10.1111/j.1745-493X.2009.03166.x(2009)

    work page doi:10.1111/j.1745-493x.2009.03166.x(2009 2009

  5. [5]

    & Lafond, F.Firm-Level Production Networks: What Do We (Really) Know?Unpublished

    Bacilieri, A., Borsos, A., Astudillo-Est´ evez, P., Hoefer, M. & Lafond, F.Firm-Level Production Networks: What Do We (Really) Know?Unpublished. 2025

    work page 2025

  6. [6]

    T.et al.Defining Supply Chain Management.Journal of Business Logistics22,1–25

    Mentzer, J. T.et al.Defining Supply Chain Management.Journal of Business Logistics22,1–25. doi:10.1002/j.2158-1592.2001.tb00001.x(2001)

    work page doi:10.1002/j.2158-1592.2001.tb00001.x(2001 2001

  7. [7]

    G., Chaddad, F

    Lazzarini, S. G., Chaddad, F. R. & Cook, M. L. Integrating Supply Chain and Network Analyses: The Study of Netchains.Journal on Chain and Network Science1,7–22. doi:10.3920/JCNS2001. x002(Mar. 2001)

    work page doi:10.3920/jcns2001 2001

  8. [8]

    & Dolgui, A

    Ivanov, D. & Dolgui, A. Viability of Intertwined Supply Networks: Extending the Supply Chain Resilience Angles towards Survivability. A Position Paper Motivated by COVID-19 Outbreak.In- ternational journal of production research58,2904–2915 (2020)

    work page 2020

  9. [9]

    Handfield, R. B. & Nichols Jr., E. L.Introduction to Supply Chain Management1999

    work page

  10. [10]

    M., Cooper, M

    Lambert, D. M., Cooper, M. C. & Pagh, J. D. Supply Chain Management: Implementation Issues and Research Opportunities.The International Journal of Logistics Management9,1–20. doi:10. 1108/09574099810805807(July 1998). 11

    work page 1998

  11. [11]

    Caddy, I. N. & Helou, M. M. Supply Chains and Their Management: Application of General Systems Theory.Journal of Retailing and Consumer Services14,319–327. doi:10.1016/j.jretconser. 2006.12.001(Sept. 2007)

    work page doi:10.1016/j.jretconser 2006

  12. [12]

    Y., Dooley, K

    Choi, T. Y., Dooley, K. J. & Rungtusanatham, M. Supply Networks and Complex Adaptive Sys- tems: Control versus Emergence.Journal of Operations Management19,351–366. doi:10.1016/ S0272-6963(00)00068-1(May 2001)

    work page 2001

  13. [13]

    Burt, R. S. The Network Structure Of Social Capital.Research in Organizational Behavior22, 345–423. doi:10.1016/S0191-3085(00)22009-1(Jan. 2000)

    work page doi:10.1016/s0191-3085(00)22009-1(jan 2000

  14. [14]

    Carvalho, V. M. & Tahbaz-Salehi, A. Production Networks: A Primer.Annual Review of Economics 11,635–663. doi:10.1146/annurev-economics-080218-030212(Aug. 2019)

    work page doi:10.1146/annurev-economics-080218-030212(aug 2019

  15. [15]

    & Aoyama, H

    Fujiwara, Y. & Aoyama, H. Large-Scale Structure of a Nation-Wide Production Network.The European Physical Journal B77,565–580. doi:10.1140/epjb/e2010-00275-2(Oct. 2010)

    work page doi:10.1140/epjb/e2010-00275-2(oct 2010

  16. [16]

    K., Mogstad, M

    Dhyne, E., Kikkawa, A. K., Mogstad, M. & Tintelnot, F. Trade and Domestic Production Networks. The Review of Economic Studies88,643–668. doi:10.1093/restud/rdaa062(Mar. 2021)

    work page doi:10.1093/restud/rdaa062(mar 2021

  17. [17]

    doi:10.1126/science.adi7521(2023)

    Pichler, A.et al.Building an Alliance to Map Global Supply Networks.Science382,270–272. doi:10.1126/science.adi7521(2023)

    work page doi:10.1126/science.adi7521(2023 2023

  18. [18]

    M., Nirei, M., Saito, Y

    Carvalho, V. M., Nirei, M., Saito, Y. U. & Tahbaz-Salehi, A. Supply Chain Disruptions: Evidence from the Great East Japan Earthquake*.The Quarterly Journal of Economics136,1255–1321. doi:10.1093/qje/qjaa044(May 2021)

    work page doi:10.1093/qje/qjaa044(may 2021

  19. [19]

    & Todo, Y

    Inoue, H. & Todo, Y. Firm-Level Propagation of Shocks through Supply-Chain Networks.Nature Sustainability2,841–847. doi:10.1038/s41893-019-0351-x(Sept. 2019)

    work page doi:10.1038/s41893-019-0351-x(sept 2019

  20. [20]

    & Verhetsel, A

    Beckers, J., Thomas, I., Vanoutrive, T. & Verhetsel, A. Logistics Clusters, Including Inter-Firm Re- lations through Community Detection.European Journal of Transport and Infrastructure Research 18.doi:10.18757/ejtir.2018.18.2.3229(Apr. 2018)

    work page doi:10.18757/ejtir.2018.18.2.3229(apr 2018

  21. [21]

    D., Guillaume, J.-L., Lambiotte, R

    Blondel, V. D., Guillaume, J.-L., Lambiotte, R. & Lefebvre, E. Fast Unfolding of Communities in Large Networks.Journal of Statistical Mechanics: Theory and Experiment2008,P10008. doi:10. 1088/1742-5468/2008/10/P10008(Oct. 2008)

    work page 2008

  22. [22]

    & Griffis, S

    Wiedmer, R. & Griffis, S. E. Structural Characteristics of Complex Supply Chain Networks.Journal of Business Logistics42,264–290 (2021)

    work page 2021

  23. [23]

    Chakraborty, A.et al.Hierarchical Communities in the Walnut Structure of the Japanese Produc- tion Network.PloS one13,e0202739 (2018)

    work page 2018

  24. [24]

    & Bergstrom, C

    Rosvall, M. & Bergstrom, C. T. Maps of random walks on complex networks reveal community structure.Proceedings Of The National Academy Of Sciences Of The United States Of America 105,1118–1123. doi:10.1073/pnas.0706851105(2008)

    work page doi:10.1073/pnas.0706851105(2008 2008

  25. [25]

    & Inoue, H

    Kichikawa, Y., Iyetomi, H., Iino, T. & Inoue, H. Community Structure Based on Circular Flow in a Large-Scale Transaction Network.Applied Network Science4.doi:10.1007/s41109-019-0202-8 (2019)

    work page doi:10.1007/s41109-019-0202-8 2019

  26. [26]

    de Jonge, E., Pijpers, F. P. & Mandjes, M. Deriving Production Chains Using Restricted Gradient Extraction.Chaos: An Interdisciplinary Journal of Nonlinear Science35.doi:10.1063/5.0270180 (2025)

    work page doi:10.1063/5.0270180 2025

  27. [27]

    & Dong, Z

    Lu, Z. & Dong, Z. A Gravitation-Based Hierarchical Community Detection Algorithm for Structur- ing Supply Chain Network.International Journal of Computational Intelligence Systems16,110 (2023). 28.TEIKOKU DATABANK, Ltd

    work page 2023

  28. [28]

    & Takayasu, M

    Tamura, K., Takayasu, H. & Takayasu, M. Extraction of conjugate main-stream structures from a complex network flow.Phys. Rev. E91,042815. doi:10.1103/PhysRevE.91.042815(4 Apr. 2015)

    work page doi:10.1103/physreve.91.042815(4 2015

  29. [29]

    & Takayasu, M

    Tamura, K., Takayasu, H. & Takayasu, M. Diffusion-localization transition caused by nonlinear transport on complex networks.Scientific reports8,5517. doi:10 . 1038 / s41598 - 018 - 23675 - x (2018). 12

    work page 2018

  30. [30]

    & Matsuo, K.PatentJP B9 6860731 (2021)

    Kimura, K., Kitajima, S., Yauchi, H. & Matsuo, K.PatentJP B9 6860731 (2021)

    work page 2021

  31. [31]

    & Kitajima, S.PatentJP 7097500 B1 (2022)

    Kimura, K., Kawabata, A., Yasue, N., Tokunaga, A. & Kitajima, S.PatentJP 7097500 B1 (2022)

    work page 2022

  32. [32]

    H., Yao, Y

    Jiang, X., Lim, L.-H. H., Yao, Y. & Ye, Y. Statistical Ranking and Combinatorial Hodge Theory. Mathematical Programming127,203–244. doi:10.1007/s10107-010-0419-x(2011)

    work page doi:10.1007/s10107-010-0419-x(2011 2011

  33. [33]

    & Fujiki, Y

    Haruna, T. & Fujiki, Y. Hodge Decomposition of Information Flow on Small-World Networks. Frontiers in Neural Circuits10.doi:10.3389/fncir.2016.00077(Feb. 2016)

    work page doi:10.3389/fncir.2016.00077(feb 2016

  34. [34]

    Yamagishi, N

    Aoki, T., Fujishima, S. & Fujiwara, N. Urban Spatial Structures from Human Flow by Hodge– Kodaira Decomposition.Scientific Reports12,11258. doi:10.1038/s41598- 022- 15512- z(July 2022)

    work page doi:10.1038/s41598- 2022

  35. [35]

    Choi, T. Y. & Kim, Y. Structural Embeddedness and Supplier Management: A Network Perspective. Journal of supply chain management44,5–13 (2008). 13 A Industrial and geographic characteristics of extracted clusters In the main text, we have summarized the industrial and geographical characteristics for the largest three clusters. In this section, we summari...

    work page 2008