Pith. sign in

REVIEW 4 major objections 6 minor 44 references

A graph neural network based on feature network for identifying influential nodes

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A GCN that selects node features by clustering correlated local centralities beats global-centrality baselines on influential-node detection.

desk verdict The feature-network selection idea is real but the paper's own ablation undermines the claim that it beats just using all local centralities. read the letter →

arxiv 2508.01278 v1 pith:3T5RYWFP submitted 2025-08-02 cs.SI cs.LG

Yanmei Hu , Siyuan Yin , Yihang Wu , Xue Yue , Yue Liu This is my paper · ORCID
classification cs.SIcs.LG
keywords influentialnodeidentificationgraphconvolutionalnetworkfeaturelocalcentralitySpearmancorrelationSIRmodelclassificationcomplexnetworks
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper tries to show that the choice of node features is the limiting factor in graph-neural-network methods for finding influential nodes, and that a principled feature-selection step fixes it. FNGCN builds a small "feature network" over eleven local centrality measures, connects centralities whose Spearman correlation exceeds a threshold, clusters them, and picks one representative from each cluster to feed a GCN. Against SVM, logistic regression, InfGCN, GATv2, and GATv2-FN on six real-world networks, both the shallow and deep FNGCN variants achieve higher accuracy, F1, and AUC in most comparisons, and they spend far less time constructing features because global centralities are avoided. The paper also concludes that local centralities alone are sufficient, that the number of GCN layers matters little, and that Conductance of Egonet is the most informative feature.

What carries the argument

The feature network is the load-bearing object: each node is a local centrality, and an edge is placed between two centralities when their Spearman correlation coefficient is above $\delta = 0.9$. Community detection on this small graph groups redundant centralities, and the selection rule keeps one centrality per group — the node with largest degree that has no already-chosen neighbor — giving a small feature set that is then rank-normalized into the node-feature matrix. The GCN receives this matrix together with the adjusted transition matrix $\tilde{P} = \tilde{D}^{-1/2}\tilde{A}\tilde{D}^{-1/2}$; the shallow variant stacks three layers with the standard convolution, while the deep variant uses the residual and identity-mapping update of [13] to stack 64 layers without over-smoothing.

What would settle it

A decisive test would run the selection with several thresholds $\delta$ and several community-detection algorithms on the same six networks and same train/test splits; if accuracy and F1 vary strongly with those choices, or if a GCN fed all eleven local centralities matches or beats the selected subset, then the feature-network selection is not carrying the reported gains.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that representing the relationships among local centralities as a feature network and using that network to select a compact, de-correlated feature set improves a GCN's ability to identify influential nodes. The experiments show FNGCN3 and FNGCN64 outperforming the comparison methods on most of the six networks across accuracy, F1, and AUC, with the main degradations confined to CA-GrQc and CA-HepTh; ablation experiments are then used to argue that local centralities are sufficient, that deep layers give only marginal gains, that centrality selection is necessary, and that Conductance of Egonet contributes most to the model.

Load-bearing premise

The load-bearing premise is that Algorithm 1's feature selection — the fixed Spearman threshold, the unspecified community-detection step, and the arbitrary tie-breaking — actually picks out the best local centralities, so that the reported accuracy gains come from the feature network rather than from the GCN itself or from per-network tuning.

Editorial extensions

If this is right

  • Feature construction drops from tens or hundreds of seconds to under one second on the six networks, because only local centralities are computed, so the method is practical on larger graphs.
  • A 3-layer GCN with selected local features is within about one point of a 64-layer GCN on most datasets, meaning expensive deep stacks are not required for this task.
  • Because the selected feature set changes from network to network (for instance, Degree is added for Human protein (Vidal) but not the others), the feature-network step is doing genuine per-network selection rather than applying a fixed list.
  • Conductance of Egonet is the single most important feature in the ablation, so future centrality-based classifiers for these networks should include it.
  • Local centralities alone match or beat global centralities in the comparisons, so the heavy cost of betweenness and closeness is avoidable.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial: because the selection step operates only on the input features and the network structure, the same framework could be dropped onto other GNN architectures or other node-classification problems; the paper only tests GCN.
  • Editorial: the fixed threshold $\delta=0.9$ and the unspecified community-detection algorithm leave room for sensitivity analysis; sweeping the threshold would clarify how much of the gain comes from the particular selection rule.
  • Editorial: the prominence of Conductance of Egonet raises the possibility that a much simpler classifier using only that centrality plus degree would already achieve most of the accuracy, a baseline the paper does not report.
  • Editorial: because node features are rank-normalized, the method should transfer across networks of different sizes without rescaling; testing on directed or weighted networks would reveal whether the local-centrality definitions remain sufficient.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 6 minor

Summary. The paper proposes FNGCN, a graph convolutional network framework for identifying influential nodes. It computes 11 local centrality features, builds a 'feature network' in which centralities are connected by Spearman correlation above a threshold, groups centralities via community detection, and selects one representative from each group as node features. Two GCN variants, shallow FNGCN3 and deep FNGCN64, are trained on SIR-model labels and evaluated on six real-world networks against SVM, LR, InfGCN, GATv2, and GATv2-FN. The paper also reports ablations on network depth, local versus global centralities, the benefit of centrality selection, and per-feature contributions. The central claim is that FNGCN identifies influential nodes more accurately than the compared methods while spending much less time on feature construction.

Significance. If the claims were established, the contribution would be moderate but useful: a GNN-based influence-identification method that uses only local centralities and a principled redundancy-selection step could reduce computational cost and improve interpretability. The manuscript has several good ingredients: the SIR ground truth is computed independently of the model, the ablation suite is broad, and the time-consumption comparison in Table VI supports the efficiency advantage of local centralities. However, the central comparative claim is not currently supported because the feature-selection algorithm is underspecified, the paper's own ablation does not consistently favor selection, and all numerical comparisons lack variance estimates. The work would be strengthened by a fully specified selection procedure, a sensitivity analysis, and controlled re-runs of all baselines under the same protocol.

major comments (4)
  1. [Section III-B (Algorithm 1)] Algorithm 1 is the only novel component separating FNGCN from a GCN with local-centrality features, but it is not fully specified. Step S3 says 'any community detection algorithm' without naming or parameterizing one; the Spearman threshold δ=0.9 is fixed with no sensitivity analysis; and step S4 breaks ties arbitrarily, as the text says 'we can choose either of them.' Because the selected feature sets differ across the six networks (Table III), the method is not deterministic or reproducible as presented. Please specify the community detection algorithm, state how δ was chosen, and report the stability of the selected features and downstream results with respect to δ and tie-breaking.
  2. [Section IV-D, Fig. 5] The paper's own ablation contradicts the claim that feature-network selection is beneficial. In Fig. 5(a), FNGCN3 is below GCN3+LC on CA-HepTh in accuracy and F1; in Fig. 5(b), FNGCN64 is below GCN64+LC on Human protein (Vidal) and CA-GrQc. Since GCN+LC uses all 11 local centralities as features, the selection step does not consistently beat simply using no selection. This undermines the attribution of the Table V gains to the feature network. The authors should either provide a systematic comparison showing when selection helps or hurts and why, or qualify the contribution as comparable to using all local centralities with reduced feature dimensionality.
  3. [Section IV-B/C, Tables V and Figs. 3-7] All reported accuracies, F1 scores, and AUC values are single point estimates, with no standard deviations, number of random seeds, or significance tests. Many differences are extremely small, on the order of 0.1 to 0.5 percentage points, so they could easily be noise. This is particularly problematic because the headline claim is that FNGCN identifies influential nodes 'more accurately' than the compared methods. Please rerun each configuration with multiple random seeds, report mean and standard deviation, and where possible include paired significance tests.
  4. [Section IV-C, Table V] The InfGCN results are transcribed from the original paper rather than re-run under the same training, labeling, and sampling pipeline, and the footnote to Table V concedes that accuracy values and all Sister-Cities results are not available from the original paper. Because FNGCN's SIR labels, 1:2 positive-negative sampling, and 70/30 train-test split may differ from the InfGCN setup, this baseline comparison is not controlled. Please re-run InfGCN under the protocol used for the other methods, or explicitly label the comparison as 'as reported in [6]' and avoid claiming superiority over it.
minor comments (6)
  1. [Throughout] There are numerous typographical errors, including 'indentifying' in the Abstract, 'redundency', 'Condctance', 'nosies', 'aviod', 'baesd', and 'diffrent'; the manuscript needs a careful copyedit.
  2. [Table I] Table I is heavily garbled by LaTeX and many formulas are unreadable, especially those for CoredCosine, CoredJaccard, and CoredPearson; the table should be regenerated so that every centrality definition is legible.
  3. [Eq. (4)] Equation (4), the Spearman correlation formula, is typeset incorrectly with misplaced parentheses and square roots, and the definitions of x_i and y_i should be stated more clearly; the current text uses 'ordinal number' but it is not clear whether ties are handled.
  4. [Reference [39]] Reference [39] contains the placeholder 'http://www.address of you wannar cite/' and should be replaced with the proper KONECT dataset citations.
  5. [Section IV-D and figure captions] The ablation numbering is inconsistent: Fig. 4's caption calls the local/global/all-centrality comparison the 'first ablation experiments,' while Section IV-D describes the hidden-layer study as the first ablation and later refers to Fig. 5 as the 'second ablation experiment'; the labels should be aligned.
  6. [Section V] The conclusion repeats the sentence 'identify the nodes with the best centrality values as influential nodes' twice; the duplicate phrase should be removed.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SIR ground truth is external and feature selection is label-free, so the evaluation does not reduce to its inputs.

full rationale

The derivation chain is self-contained and not circular. The SIR influence labels (Eq. 6) are computed from independent stochastic simulations and never enter Algorithm 1's feature selection, which uses only Spearman correlations among local centralities and the feature-network topology; no target labels are used to choose features, so the reported accuracy/F1/AUC are measured against an external ground truth rather than reconstructed from the inputs. The only overlap with prior work is the motivational citation [22] (author Hu et al.), which is used to support the general idea that relationships among features matter; Algorithm 1 is specified in this paper and its validity is not reduced to that citation. The underspecified threshold δ=0.9, unnamed community detection, and the ablation results in Fig. 5 where FNGCN is sometimes below GCN+LC are reproducibility/correctness concerns, not circularity: they challenge whether the claimed benefit of feature-network selection is established, but they do not make any prediction equivalent to the input by construction.

Assumptions & free parameters 7 free parameters · 5 assumptions · 1 invented entities

The central claim rests on the SIR-based labeling, a fixed Spearman threshold δ=0.9, an unspecified community detection step, and a greedy representative selection with arbitrary tie-breaking. The SIR assumption is standard in this literature, but the threshold, the clustering method, and the tie-breaks are ad hoc choices that are never validated by sensitivity analysis. The GCN and GCNII propagation equations are standard background math. No physical entities are postulated; the only constructed object is the feature network itself, which is reproducible in principle but not fully specified.

free parameters (7)
  • δ (Spearman correlation threshold) = 0.9
    Fixed in Section III-B for all networks with no sensitivity analysis; the chosen centralities in Table III and hence all results depend on it.
  • Proportion of nodes labeled influential = Top 5% of SIR ranking
    Section III-C defines influential nodes as the top 5% by SIR score; this cutoff defines the classification task, is taken from prior work [7], and is never varied.
  • Negative sampling scheme = 10% of non-influential nodes, 1:2 positive:negative
    Section IV-B samples a small fraction of the 95% negative pool; accuracy and F1 are computed on this subsample, not on the full network.
  • SIR infection and recovery rates = β=2βc, γ=1
    Section III-C Eq. (7); follows [7], but the ground-truth labels change with β and no other β is tested.
  • Number of SIR runs per node = 1000
    Section III-C Eq. (6); chosen to reduce stochasticity, standard practice.
  • Community detection algorithm (Algorithm 1 step S3) = Unspecified
    The text says 'any community detection algorithm' can be used but never names the one that produced Table III, so the feature selection is not reproducible.
  • Tie-breaking in step S4 = Arbitrary ('we can choose either of them')
    For CA-CondMat, the choice between CoredCosine and CoredPearson is arbitrary; the reported results depend on it.
assumptions (5)
  • domain assumption SIR spread size (with β=2βc, γ=1, averaged over 1000 runs) is the correct notion of node influence used as ground truth
    The whole supervised training and evaluation relies on SIR-based labels (Section III-C); if influence is instead defined by other dynamics, the ranking and the conclusions could change.
  • ad hoc to paper Spearman correlation above 0.9 between two centralities means they are redundant and one can represent the group
    The threshold and the correlation measure are the core of Algorithm 1 (Section III-B, Eq. 4) but are not derived or validated; no ablation varies δ.
  • domain assumption The mean-field epidemic threshold formula βc = <d>/(<d^2> - <d>)
    Used in Eq. (7), Section III-C, following [7] to set the infection rate; this formula assumes locally tree-like structure that real networks violate.
  • ad hoc to paper Community structure in the feature network corresponds to groups of mutually redundant centralities
    Step S3 of Algorithm 1 clusters centrality nodes; the mapping from detected communities to redundant feature groups is assumed without evidence, and the algorithm is unspecified.
  • standard math Standard GCN and GCNII propagation rules
    The model updates H^{l+1} = σ(P̃ H^l W^l) for FNGCN3 and the GCNII residual/identity variant for FNGCN64, both from refs [14] and [13].
invented entities (1)
  • Feature network independent evidence
    purpose: A graph whose nodes are centrality measures and whose edges connect centralities with Spearman correlation above threshold δ; used to cluster and then select one representative centrality per group as GCN node features (Algorithm 1, Section III-B).
    This is a methodological data structure, not a physical entity. It is in principle reconstructible from the public datasets, but full verifiability is limited because the community detection algorithm (step S3) and tie-breaking (step S4) are unspecified.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A graph neural network based on feature network for identifying influential nodes." pith.science (2026). https://pith.science/paper/3T5RYWFP

@misc{pith2026250801278,
  author       = {Pith},
  title        = {Pith review of: A graph neural network based on feature network for identifying influential nodes},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/3T5RYWFP}},
  note         = {Machine review of arXiv:2508.01278}
}
read the original abstract

Identifying influential nodes in complex networks is of great importance, and has many applications in practice. For example, finding influential nodes in e-commerce network can provide merchants with customers with strong purchase intent; identifying influential nodes in computer information system can help locating the components that cause the system break down and identifying influential nodes in these networks can accelerate the flow of information in networks. Thus, a lot of efforts have been made on the problem of indentifying influential nodes. However, previous efforts either consider only one aspect of the network structure, or using global centralities with high time consuming as node features to identify influential nodes, and the existing methods do not consider the relationships between different centralities. To solve these problems, we propose a Graph Convolutional Network Framework based on Feature Network, abbreviated as FNGCN (graph convolutional network is abbreviated as GCN in the following text). Further, to exclude noises and reduce redundency, FNGCN utilizes feature network to represent the complicated relationships among the local centralities, based on which the most suitable local centralities are determined. By taking a shallow GCN and a deep GCN into the FNGCN framework, two FNGCNs are developed. With ground truth obtained from the widely used Susceptible Infected Recovered (SIR) model, the two FNGCNs are compared with the state-of-art methods on several real-world networks. Experimental results show that the two FNGCNs can identify the influential nodes more accurately than the compared methods, indicating that the proposed framework is effective in identifying influential nodes in complex networks.

Figures

Figures reproduced from arXiv: 2508.01278 by the authors.

Figure 1
Figure 1. The framework of FNGCN. The input layer prepares [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. The feature network of CA-CondMat TABLE III: Local centralities for each network Network Centrality Hamsterster Friends,Sister-Cities, CA-CondMat SPA, CoredCosine, CoredPearson, Density of Egonet,Conductance, LCC Human protein (Vidal) SPA, CoredCosine, CoredPearson, Density of Egonet, Conductance, LCC, Degree CA-GrQc, CA-HepTh SPA, CoredCosine, Density of Egonet, Conductance, LCC C. SIR model Similar to other GNNs u… view at source ↗
Figure 3
Figure 3. The results of the impact of the number of hidden GCN [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The results of the first ablation experiments, i.e., [PITH_FULL_IMAGE:figures/full_fig_p008_4.png]
Figure 7
Figure 7. Figure 7: The contribution of different local centralities based on [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]
Figure 6
Figure 6. Figure 6: The contribution of different local centralities based on [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

44 extracted references · 40 canonical work pages

  1. [39]

    “Konect,” 2017, http://www.address of you wannar cite/

    work page 2017

  2. [6]

    Infgcn: Identifying influential nodes in complex networks with graph convolutional networks,

    G. Zhao, P. Jia, A. Zhou, and B. Zhang, “Infgcn: Identifying influential nodes in complex networks with graph convolutional networks,” Neuro- computing, vol. 414, pp. 18–26, 2020

    work page 2020

  3. [1]

    Graph neural networks in recommender systems: A survey,

    S. Wu, F. Sun, W. Zhang, X. Xie, and B. Cui, “Graph neural networks in recommender systems: A survey,” ACM Comput. Surv., vol. 55, no. 5, 2022

    work page 2022

  4. [2]

    A novel method for analog fault diagnosis based on neural networks and genetic algorithms,

    Y . Tan, Y . He, C. Cui, and G. Qiu, “A novel method for analog fault diagnosis based on neural networks and genetic algorithms,” IEEE Transactions on Instrumentation and Measurement , vol. 57, no. 11, pp. 2631–2639, 2008

    work page 2008

  5. [3]

    Table structure recognition and form parsing by end-to-end object detection and relation parsing,

    X.-H. Li, F. Yin, H.-S. Dai, and C.-L. Liu, “Table structure recognition and form parsing by end-to-end object detection and relation parsing,” Pattern Recognition, vol. 132, p. 108946, 2022

    work page 2022

  6. [4]

    Identifying spreading influence nodes for social networks,

    O. Y . G. Q. L. Jianguo, “Identifying spreading influence nodes for social networks,” Frontiers of Engineering Management , vol. 9, pp. 520–549, 2022

    work page 2022

  7. [5]

    Identifying critical nodes in complex networks via graph convolutional networks,

    E.-Y . Yu, Y .-P. Wang, Y . Fu, D.-B. Chen, and M. Xie, “Identifying critical nodes in complex networks via graph convolutional networks,” Knowledge-Based Systems , vol. 198, p. 105893, 2020

    work page 2020

  8. [7]

    A new approach for evaluating node importance in complex networks via deep learning methods,

    M. Zhang, X. Wang, L. Jin, M. Song, and Z. Li, “A new approach for evaluating node importance in complex networks via deep learning methods,” Neurocomputing, vol. 497, pp. 13–27, 2022

    work page 2022

Show all 44 references
  1. [8]

    Centrality indices,

    D. Kosch ¨utzki, K. A. Lehmann, L. Peeters, S. Richter, D. Tenfelde- Podehl, and O. Zlotowski, “Centrality indices,” Network analysis: methodological foundations , pp. 16–61, 2005

    2005

  2. [9]

    A model of internet topology using ¡i¿k¡/i¿-shell decomposition,

    S. Carmi, S. Havlin, S. Kirkpatrick, Y . Shavitt, and E. Shir, “A model of internet topology using ¡i¿k¡/i¿-shell decomposition,” Proceedings of the National Academy of Sciences , vol. 104, no. 27, pp. 11 150–11 154, 2007

    2007

  3. [10]

    Support vector machines,

    M. Hearst, S. Dumais, E. Osuna, J. Platt, and B. Scholkopf, “Support vector machines,” IEEE Intelligent Systems and their Applications , vol. 13, no. 4, pp. 18–28, 1998

    1998

  4. [11]

    Logistic regression,

    M. P. LaValley, “Logistic regression,” Circulation, vol. 117, no. 18, pp. 2395–2399, 2008

    2008

  5. [12]

    Graph neural networks: A review of methods and applications,

    J. Zhou, G. Cui, S. Hu, Z. Zhang, and C. Yang, “Graph neural networks: A review of methods and applications,”AI Open, vol. 1, pp. 57–81, 2020

    2020

  6. [13]

    Simple and deep graph convolutional networks,

    M. Chen, Z. Wei, Z. Huang, B. Ding, and Y . Li, “Simple and deep graph convolutional networks,” in International conference on machine learning. PMLR, 2020, pp. 1725–1735

    2020

  7. [14]

    Semi-supervised classification with graph convolutional networks,

    T. N. Kipf and M. Welling, “Semi-supervised classification with graph convolutional networks,” arXiv preprint arXiv:1609.02907 , 2016

    2016 arXiv

  8. [15]

    Mathematical modeling of diseases: Susceptible-infected-recovered (sir) model,

    T. Johnson and B. McQuarrie, “Mathematical modeling of diseases: Susceptible-infected-recovered (sir) model,” University of Minnesota, Morris, Math , vol. 4901, 2009

    2009

  9. [16]

    Ranking of closeness centrality for large-scale social networks,

    K. Okamoto, W. Chen, and X.-Y . Li, “Ranking of closeness centrality for large-scale social networks,” in International workshop on frontiers in algorithmics . Springer, 2008, pp. 186–195

    2008

  10. [17]

    Betweenness centrality in large complex networks,

    M. Barthelemy, “Betweenness centrality in large complex networks,” The European physical journal B , vol. 38, no. 2, pp. 163–168, 2004

    2004

  11. [18]

    A. N. Langville and C. D. Meyer, Google’s PageRank and beyond: The science of search engine rankings . Princeton university press, 2006

    2006

  12. [19]

    Eigenvector-centrality—a node-centrality?

    B. Ruhnau, “Eigenvector-centrality—a node-centrality?” Social net- works, vol. 22, no. 4, pp. 357–365, 2000. JOURNAL OF LATEX CLASS FILES, VOL. 18, NO. 9, SEPTEMBER 2020 11

    2000

  13. [20]

    Graph convolutional networks: a comprehensive review,

    S. Zhang, H. Tong, J. Xu, and R. Maciejewski, “Graph convolutional networks: a comprehensive review,” Computational Social Networks , vol. 6, no. 1, pp. 1–23, 2019

    2019

  14. [21]

    Fg-rs: Capture user fine-grained preferences through attribute information for recommender systems,

    H. Chen, F. Qian, J. Chen, S. Zhao, and Y . Zhang, “Fg-rs: Capture user fine-grained preferences through attribute information for recommender systems,” Neurocomputing, vol. 458, pp. 195–203, 2021

    2021

  15. [22]

    Differential evolution based on network structure for feature selection,

    Y . Hu, M. Lu, X. Li, and B. Cai, “Differential evolution based on network structure for feature selection,” Information Sciences, vol. 635, pp. 279–297, 2023

    2023

  16. [23]

    A multi-level damage classification technique of aircraft plate structures using lamb wave- based deep transfer learning network,

    W. Shao, H. Sun, Y . Wang, and X. Qing, “A multi-level damage classification technique of aircraft plate structures using lamb wave- based deep transfer learning network,” Smart Materials and Structures , vol. 31, no. 7, p. 075019, 2022

    2022

  17. [24]

    A novel flexible sensor for double- parameter decoupling measurement of temperature and pressure with high sensitivity and wide range,

    L. Fan, X. Yang, and H. Sun, “A novel flexible sensor for double- parameter decoupling measurement of temperature and pressure with high sensitivity and wide range,” Journal of Materials Chemistry C , vol. 11, no. 30, pp. 10 163–10 177, 2023

    2023

  18. [25]

    On node features for graph neural networks,

    C. T. Duong, T. D. Hoang, H. T. H. Dang, Q. V . H. Nguyen, and K. Aberer, “On node features for graph neural networks,” arXiv preprint arXiv:1911.08795, 2019

    1911 arXiv

  19. [26]

    Comparative analysis of centrality measures for identifying critical nodes in complex networks,

    O. Ugurlu, “Comparative analysis of centrality measures for identifying critical nodes in complex networks,” Journal of Computational Science , vol. 62, p. 101738, 2022

    2022

  20. [27]

    Ranking nodes in complex networks based on local structure and improving closeness centrality,

    C. Salavati, A. Abdollahpouri, and Z. Manbari, “Ranking nodes in complex networks based on local structure and improving closeness centrality,” Neurocomputing, vol. 336, pp. 36–45, 2019

    2019

  21. [28]

    Acquiring academic literacy: a case of first-year extended degree programme students,

    S. Van Schalkwyk, E. Bitzer, and C. van der Walt, “Acquiring academic literacy: a case of first-year extended degree programme students,” Southern African Linguistics and Applied Language Studies , vol. 27, no. 2, pp. 189–201, 2010

    2010

  22. [29]

    Using accumulated degree-days to estimate the postmortem interval from decomposed human remains,

    M. S. Megyesi, S. P. Nawrocki, and N. H. Haskell, “Using accumulated degree-days to estimate the postmortem interval from decomposed human remains,” Journal of forensic sciences , vol. 50, no. 3, p. JFS2004017, 2005

    2005

  23. [30]

    A robust two-stage algorithm for local community detection,

    X. Ding, J. Zhang, and J. Yang, “A robust two-stage algorithm for local community detection,” Knowledge-Based Systems , vol. 152, pp. 188– 199, 2018

    2018

  24. [31]

    Overlapping community detection using neighborhood-inflated seed expansion,

    J. J. Whang, D. F. Gleich, and I. S. Dhillon, “Overlapping community detection using neighborhood-inflated seed expansion,” IEEE Transac- tions on Knowledge and Data Engineering , vol. 28, no. 5, pp. 1272– 1284, 2016

    2016

  25. [32]

    Local clustering coeffi- cient in generalized preferential attachment models,

    A. Krot and L. Ostroumova Prokhorenkova, “Local clustering coeffi- cient in generalized preferential attachment models,” in Algorithms and Models for the Web Graph: 12th International Workshop, WA W 2015, Eindhoven, The Netherlands, December 10-11, 2015, Proceedings 12 . Sprin...

    2015

  26. [33]

    Community detection in complex networks by detecting and expanding core nodes through ex- tended local similarity of nodes,

    K. Berahmand, A. Bouyer, and M. Vasighi, “Community detection in complex networks by detecting and expanding core nodes through ex- tended local similarity of nodes,” IEEE Transactions on Computational Social Systems , vol. 5, no. 4, pp. 1021–1033, 2018

    2018

  27. [34]

    Node classification with graph neural network based centrality measures and feature selection,

    A. M. Mahmoud, A. S. Desuky, H. F. Eid, and H. A. Ali, “Node classification with graph neural network based centrality measures and feature selection,” International Journal of Electrical and Computer Engineering, vol. 13, no. 2, p. 2114, 2023

    2023

  28. [35]

    Spearman correlation coefficients, differ- ences between,

    L. Myers and M. J. Sirois, “Spearman correlation coefficients, differ- ences between,” Encyclopedia of statistical sciences , vol. 12, 2004

    2004

  29. [36]

    Similarity detection method of science fiction painting based on multi-strategy improved sparrow search algorithm and gaussian pyramid,

    G. Chen, D. Zhu, and X. Chen, “Similarity detection method of science fiction painting based on multi-strategy improved sparrow search algorithm and gaussian pyramid,” Multimedia Tools and Applications , pp. 1–40, 2023

    2023

  30. [37]

    A sir model assumption for the spread of covid-19 in different communities,

    I. Cooper, A. Mondal, and C. G. Antonopoulos, “A sir model assumption for the spread of covid-19 in different communities,” Chaos, Solitons & Fractals, vol. 139, p. 110057, 2020

    2020

  31. [38]

    Sir-im: Sir rumor spreading model with influence mechanism in social networks,

    L. Qiu, W. Jia, W. Niu, M. Zhang, and S. Liu, “Sir-im: Sir rumor spreading model with influence mechanism in social networks,” Soft Computing, vol. 25, pp. 13 949–13 958, 2021

    2021

  32. [40]

    Graph evolution: Den- sification and shrinking diameters,

    J. Leskovec, J. Kleinberg, and C. Faloutsos, “Graph evolution: Den- sification and shrinking diameters,” ACM transactions on Knowledge Discovery from Data (TKDD) , vol. 1, no. 1, pp. 2–es, 2007

    2007

  33. [41]

    Adam: A method for stochastic optimization,

    D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” arXiv preprint arXiv:1412.6980 , 2014

    2014 arXiv

  34. [42]

    Thresholding classifiers to maximize f1 score,

    Z. Chase Lipton, C. Elkan, and B. Narayanaswamy, “Thresholding classifiers to maximize f1 score,” arXiv e-prints, pp. arXiv–1402, 2014

    2014

  35. [43]

    The use of the area under the roc curve in the evaluation of machine learning algorithms,

    A. P. Bradley, “The use of the area under the roc curve in the evaluation of machine learning algorithms,” Pattern recognition, vol. 30, no. 7, pp. 1145–1159, 1997

    1997

  36. [44]

    How attentive are graph attention networks?

    S. Brody, U. Alon, and E. Yahav, “How attentive are graph attention networks?” in International Conference on Learning Representations , 2022

    2022

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.