Pith. sign in

REVIEW 4 major objections 6 minor 59 references

Pseudo-Likelihood Ratio Screening based on Network Data with Applications

T0 review · 4 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

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

Pith's one-line read A pseudo-likelihood ratio statistic can screen ultra-high-dimensional categorical tag data and recover exactly the features that drive both responses and network connections.

desk verdict The central statistic as defined in (5)-(7) is not a log-likelihood ratio and the theory does not apply to it, so the paper's main claims are unsupported in the current form, though the underlying screening idea is worth developing. read the letter →

arxiv 2505.12695 v1 pith:TNQC7MSL submitted 2025-05-19 stat.ME

Wei Hu , Danyang Huang , Bo Zhang This is my paper · ORCID
classification stat.ME MSC 62F1262H1262H30
keywords featurescreeningpseudo-likelihoodrationetworkdatasureindependencestrongconsistencyultra-high-dimensionalcategoricalself-relatedfeaturesnetwork-related
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

This paper tackles feature screening for ultra-high-dimensional categorical tag data observed together with a social network, where the response (a user's preference) may be driven both by the user's own tags and by tags that shape who connects to whom. The central claim is that a single marginal statistic, the log pseudo-likelihood ratio $\Lambda_j = n^{-1}(\log L_j - \log L_0)$ comparing a network pseudo-likelihood without features to one with feature $j$, detects both 'self-related' and 'network-related' features at once. The paper proves strong screening consistency: under conditions (C1)-(C3), the selected set equals the true union of the two types with probability tending to one, with faster convergence when all signal flows through edge formation. If correct, this turns the network structure from a complication into a source of selection signal, and gives practitioners a model-free way to shrink thousands of tags to a handful before downstream classification.

What carries the argument

The central object is the log pseudo-likelihood ratio statistic $\Lambda_j = n^{-1}(\log L_j - \log L_0)$, where $L_0$ and $L_j$ are products over nodes and response levels of posterior factors derived under two working models: edges are conditionally independent given the responses (Case 1), or given the responses and the single candidate feature $X_j$ (Case 2). The statistic carries the argument because it separates into a self-related component and a network-related component, and because the screening theorem shows that the population version $\Lambda_j$ is exactly zero for irrelevant features while remaining positive for every feature in $S_T$. The conditional-independence working model is the engine that makes the statistic computable from observed frequencies, gives it tractable null chi-square asymptotics, and lets the same construction handle interactions by treating pairs of features as candidate items.

What would settle it

Generate $n=500$ nodes with $p=1000$ binary tags, draw $Y$ independent of all tags, and draw edges with probability $\text{Sigmoid}(c \cdot I(X_{i1}=X_{i2}=1))$ for two interacting tags while all other tags have zero edge effect; under the paper's (C3)-type signal condition no single tag changes the edge-generation probability marginally, so if PLR-SIS fails to select the interacting pair the claim that it recovers exactly $S_T$ in this regime is falsified.

Watch

Extended reading notes

Core claim

The paper's discovery is that feature relevance in network-involved classification can be defined by two separate mechanisms, directly changing the response distribution and changing the edge-generation probability, and that both mechanisms are picked up by one pseudo-likelihood ratio statistic. For each candidate feature $j$, the statistic $\Lambda_j = n^{-1}(\log L_j - \log L_0)$ is built from plug-in frequency estimators of $\pi^{r_1r_2}_y$ and $\pi^{r_1r_2k_1k_2}_{yj}$, and it decomposes as $\Lambda_j = n^{-1}(\Lambda_j^{\text{self}} + \Lambda_j^{\text{network}})$. Self-related features register in the first term through the difference between $\pi^{r|k}_{y|j}$ and $\pi^r_y$; network-related features register in the second through the relative change in edge probabilities. The paper establishes that $\Lambda_j = 0$ for irrelevant features and is bounded away from zero for every $j$ in $S_T$, yielding $P(\hat{S} = S_T) = 1 - \eta\exp(\log p - \mu n) \to 1$ under (C4), ranking consistency, a stronger $n^2$ rate when $S_Y \subset S_A$, interaction screening over $O(p^2)$ pairs, and null chi-square limits for the two components.

Load-bearing premise

The load-bearing premise is that a feature's full relevance is captured by the marginal pseudo-likelihood that conditions on at most the response and that one feature; if edge formation or response actually depends on combinations of features or on unobserved confounders, the screening consistency proof does not apply.

Editorial extensions

If this is right

  • PLR-SIS gives a model-free screening step for network classification: no model for $P(X,Y)$ or $P(A|X,Y)$ needs to be specified before selecting features.
  • When all true signals are network-related, exact feature recovery requires less data: the failure probability drops to $\eta\exp(\log p - \mu n^2)$, not $\eta\exp(\log p - \mu n)$.
  • Interactions among features can be screened in the same pass, so the procedure extends naturally from main-effect sets to sets that include pair products.
  • The null distributions of the self and network components are approximately chi-square, so threshold choices can be informed by p-values rather than by arbitrary cut-offs.
  • On the Sina Weibo data, the procedure selects 9 features instead of the 77 selected by a tag-only categorical screening method and gives higher prediction AUC, consistent with the claim that network-related features add predictive signal.

Reading between the lines

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

  • An implicit consequence is that PLR-SIS can be read as a collection of conditional-independence screens: if one trusts the working model, $\Lambda_j$ is a test statistic for whether $Y$ or $A$ depends on $X_j$, so the method doubles as an exploratory tool for classifying each tag's role as self-related, network-related, or both.
  • A natural stress test beyond the paper's simulations is to let edge formation depend on a low-dimensional latent community variable that correlates with many tags; the paper's conditions do not cover unobserved confounders, and one would expect the screening guarantee to degrade gracefully or fail.
  • The binomial edge assumption could be relaxed to a logistic degree-corrected model; the same pseudo-likelihood ratio construction should remain consistent if the correction is known or estimated at a fast enough rate, though the paper does not prove this.
  • Because the statistic separates self and network contributions, one could build a two-stage pipeline that screens with $\Lambda_j$ and then fits a downstream network classifier using only the selected features; the paper's Type-II/III classifier comparison suggests the gain should persist for other network-aware classifiers.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

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 PLR-SIS, a marginal feature-screening procedure for ultra-high-dimensional categorical features when the response and network edges are interdependent. It introduces self-related features, which directly affect the response, and network-related features, which affect edge formation, and constructs a log pseudo-likelihood ratio Lambda_j comparing a pseudo-likelihood without features with one that includes feature X_j. The main theoretical claims are strong screening consistency (Theorem 2), ranking consistency (Corollary 3), an improved dimension-growth rate when all true features are network-related (Corollary 4), interaction screening (Theorem 5), and chi-square null distributions for the self and network components (Proposition 6). The numerical evaluation includes nine simulation examples and a Sina Weibo application, reporting that PLR-SIS outperforms PC-SIS in feature recovery and prediction accuracy.

Significance. Network-aware feature screening for categorical data is a relevant and underdeveloped problem, and the conceptual separation of features into self-related and network-related components is a useful contribution. If the intended statistic were correctly defined, the PLR-SIS idea would give a principled marginal screening procedure with a clean self/network decomposition and a natural chi-square calibration, extending PC-SIS to network data. The paper also reports substantial prediction gains over PC-SIS on a real Weibo dataset. However, none of these strengths can be credited to the manuscript as written, because the displayed statistic is not the one analyzed in the theorems or, evidently, the one used in the simulations.

major comments (4)
  1. [Section 2.4, Eq. (7), and Algorithm 1] The literal definition of L0 and Lj as products over every node i1 and every response level r1 makes Lambda_j = n^{-1} sum_{i1} sum_{r1} log(ell*_{i1 r1}/ell_{i1 r1}). For a self-related feature with no network effect, this reduces to n^{-1} sum_i sum_r log(pi^{r|X_{ij}}_{y|j}/pi^r_y), which is not the likelihood-ratio statistic n^{-1} sum_i log(pi^{Y_i|X_{ij}}_{y|j}/pi^{Y_i}_y). The displayed quantity can be negative for informative features: with K=R=2, P(Y=1|X=1)=0.9, P(Y=1|X=0)=0.1, and P(Y=1)=0.5, every node contributes log(1.8)+log(0.2)=-1.02, so a perfectly predictive self-related feature has Lambda_j<0 and is never selected by the rule bS={j: bLambda_j>c*} with c*>0. Theorem 2 and Corollary 3 are therefore not theorems about the statistic that is defined and computed in the paper, and the simulation results in Table 3 cannot be reproducing the stated procedure unless a correction is made and reported.
  2. [Section 2.5] The statement that 'the PMLEs can be proved to be the observation frequencies' is inconsistent with the displayed pseudo-likelihood. Maximizing the self-part n sum_r log pi^r_y subject to sum_r pi^r_y = 1 gives pi^r_y = 1/R for all r, not n^r_y/n; the observed frequencies are the maximizers only for the product over observed response levels, product_i ell_{i,Y_i}. Similarly, for each level k of X_j, the factor n_k sum_r log pi^{r|k}_{y|j} is maximized by the uniform distribution over r, not by the observed conditional frequencies n^{rk}_{yj}/n^k_j. Thus the plug-in estimators used in Algorithm 1 are not the PMLEs for the objective defined in Eq. (7).
  3. [Section 3, Table 3, and Section 4] Because Algorithm 1 as written computes the product over all response levels, the reported coverage probabilities, such as CP(X1)=1.00 and CP(X2)=0.99 in Example 1, cannot be reproduced from the procedure described in the text. The real-data decomposition of bLambda_self_j and bLambda_network_j is likewise not attributable to Eq. (7). The manuscript must state the statistic actually used in the numerical work and, after correcting the definition, re-run the simulations and the real-data analysis; without this, the empirical claims are not tied to the proposed method.
  4. [Section 2.6, Proposition 6] The claimed asymptotic distributions also fail under the displayed definition. For the self component, 2 bLambda_self_j is not the usual likelihood-ratio chi-square statistic for independence, because the product over r changes both the sign and the scaling; the example in the first comment gives a negative value for a relevant feature, so it cannot converge in distribution to a chi-square random variable. The degrees of freedom (R-1)(K-1) and R^2(K^2-1) correspond to the corrected observed-level statistic, not to Eq. (7).
minor comments (6)
  1. [Section 1.3] The word 'summerized' should be 'summarized'.
  2. [Section 2.2] The definition of Y_{(-i)} contains a typographical double comma: 'i1 not equal to i,, 1 <= i1 <= n'.
  3. [Section 2.4] The term 'model-free' is overstated, since the statistic is constructed under explicit conditional independence and Bernoulli edge-generation assumptions; this should be qualified in the introduction and abstract.
  4. [Section 3.2] The referenced tables D1-D3 in Appendix D are not included in the arXiv submission; please ensure the supplementary material is complete and available in any resubmission.
  5. [Section 2 and Figure 3] The notation summary in Figure 3 is not fully explained in the text; the caption should define the tables and their row/column sums more explicitly.
  6. [Section 4, Table 5] The paired t-test p-values in Table 5 should be interpreted cautiously because the 50 replications use overlapping random 70/30 splits, so the effective sample size is smaller than 50; please state whether any correction for overlapping samples was applied.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: PLR-SIS is derived from stated pseudo-likelihood assumptions and evaluated on simulations and holdout data; the flagged Eq. (7) issue is a correctness/consistency problem, not a circular reduction.

full rationale

The derivation of the PLR statistic is self-contained: Section 2.4 builds the posterior quantities from explicitly stated conditional-independence cases (Case 1 and Case 2), and Section 2.5 replaces the population parameters by frequency estimators. No parameter is fitted to the target feature set, and the screening threshold is chosen by a standard maximum-ratio rule rather than by inverting the theorem. Condition (C3) is a beta-min style signal-separation assumption; assuming it does not make the consistency theorem tautological, because the theorem still requires concentration of the estimated statistic around its population value, which is not built into the condition. The oracle decomposition (2)-(3) is adopted as an assumption, supported by external literature, and is not a self-citation chain. Self-citations (PC-SIS, network naive Bayes) appear only as comparators or simulation generators, not as load-bearing justifications. The real-data analysis uses a train/test split with 50 replications, providing external validation. The main text defers all proofs to the supplementary appendix; while this makes verification harder, it is not circularity. A separate concern flagged during review: Eq. (7) defines L0 and Lj as products over every node i1 and every response level r1 rather than only the observed level Y_i1. Literally, this makes the self-part of Lambda_j equal to n^{-1} sum_i sum_r log(pi^{r|X_ij}_{y|j}/pi^r_y), which can be negative for a predictive feature and is not the log-likelihood-ratio statistic that would justify the claim that the PMLEs are observed frequencies. This is an internal mathematical inconsistency in the stated definition, not a circular derivation of the claimed result from its inputs, so it does not affect the circularity score.

Assumptions & free parameters 1 free parameters · 6 assumptions · 0 invented entities

The method's central claim is conditional on the pseudo-likelihood construction, the oracle decomposition, and the signal-separation and regularity conditions. The only tuning lever in the procedure is the screening cutoff c*, chosen data-dependently. No new physical or latent entities are introduced. The main unresolved issue is that the displayed product-over-all-response-levels definition of L0 and Lj is inconsistent with the claimed PMLE and chi-square results.

free parameters (1)
  • Screening cutoff c* = Data-driven via max ratio bLambda_(j)/bLambda_(j+1) in Algorithm 1; hard rule d=[n/log n] also suggested
    The estimated feature set is {j: bLambda_j > c*}; the threshold choice affects finite-sample false positives and negatives. The theory assumes a fixed c* between Lambda_(d+1) and Lambda_(d).
assumptions (6)
  • domain assumption Edges are conditionally independent given the response (Case 1) and given response plus one feature (Case 2), with Bernoulli edge probabilities.
    Used to derive the pseudo-likelihoods in Section 2.4, eqs. (5)-(6). This is a modeling assumption, not a consequence of the data.
  • domain assumption The oracle decomposition (2)-(3) holds, so ST = SA union SY is sufficient for response prediction.
    Defines the target feature set; if the network depends on features outside SA or responses depend on features outside SY in a way not captured by f, the method's target is mis-specified. Stated in Section 2.2.
  • domain assumption Signal separation condition (C3): irrelevant features have population Lambda_j = 0 and relevant features have signal bounded away from zero.
    This is the beta-min type condition that lets marginal screening separate true from null features. It is assumed, not derived from data (Section 2.6).
  • domain assumption Technical conditions (C1), (C2), (C4): bounded level probabilities, edge probabilities between kappa1 n^{-gamma} and 1 - kappa2 n^{-gamma}, and log p = O(n^xi).
    Standard regularity conditions for strong screening consistency, stated in Section 2.6 before Theorem 2.
  • domain assumption R, K, and d are fixed, and p grows at most exponentially in n^xi.
    Used for degrees of freedom in Proposition 6 and for the convergence rate in Theorem 2; stated at the start of Section 2.6.
  • standard math The supplementary proofs rely on standard concentration and empirical process inequalities.
    Not verifiable from this submission, which lacks the supplementary file.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Pseudo-Likelihood Ratio Screening based on Network Data with Applications." pith.science (2026). https://pith.science/paper/TNQC7MSL

@misc{pith2026250512695,
  author       = {Pith},
  title        = {Pith review of: Pseudo-Likelihood Ratio Screening based on Network Data with Applications},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TNQC7MSL}},
  note         = {Machine review of arXiv:2505.12695}
}
read the original abstract

Social network platforms today generate vast amounts of data, including network structures and a large number of user-defined tags, which reflect users' interests. The dimensionality of these personalized tags can be ultra-high, posing challenges for model analysis in targeted preference analysis. Traditional categorical feature screening methods overlook the network structure, which can lead to incorrect feature set and suboptimal prediction accuracy. This study focuses on feature screening for network-involved preference analysis based on ultra-high-dimensional categorical tags. We introduce the concepts of self-related features and network-related features, defined as those directly related to the response and those related to the network structure, respectively. We then propose a pseudo-likelihood ratio feature screening procedure that identifies both types of features. Theoretical properties of this procedure under different scenarios are thoroughly investigated. Extensive simulations and real data analysis on Sina Weibo validate our findings.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

59 extracted references · 54 canonical work pages

  1. [1]

    Fokianos , Konstantinos K

    barticle [author] Armillotta , Mirko M. Fokianos , Konstantinos K. ( 2023 ). Nonlinear network autoregression . The Annals of Statistics 51 2526--2552 . barticle

    work page 2023

  2. [2]

    barticle [author] Berrett , Thomas B T. B. , Wang , Yi Y. , Barber , Rina Foygel R. F. Samworth , Richard J R. J. ( 2020 ). The conditional permutation test for independence while controlling for confounders . Journal of the Royal Statistical Society Series B: Statistical Methodology 82 175--197 . barticle

    work page 2020

  3. [3]

    ( 2013 )

    bbook [author] Bruggeman , Jeroen J. ( 2013 ). Social networks: An introduction . Routledge . bbook

    work page 2013

  4. [4]

    , Guan , Guoyu G

    barticle [author] Cai , Wei W. , Guan , Guoyu G. , Pan , Rui R. , Zhu , Xuening X. Wang , Hansheng H. ( 2018 ). Network linear discriminant analysis . Computational Statistics & Data Analysis 117 32--44 . barticle

    work page 2018

  5. [5]

    , Ferraro , Carla C

    barticle [author] Campbell , Colin C. , Ferraro , Carla C. Sands , Sean S. ( 2014 ). Segmenting consumer reactions to social network marketing . European Journal of Marketing . barticle

    work page 2014

  6. [6]

    barticle [author] Chang , Jae Ho J. H. Paul , Subhadeep S. ( 2024 ). Embedding Network Autoregression for time series analysis and causal peer effect inference . arXiv preprint arXiv:2406.05944 . barticle

    arXiv 2024

  7. [7]

    , Johansson , Fredrik D F

    barticle [author] Chen , Irene I. , Johansson , Fredrik D F. D. Sontag , David D. ( 2018 ). Why is my classifier discriminatory? Advances in Neural Information Processing Systems 31 . barticle

    work page 2018

  8. [8]

    , Li , Runze R

    barticle [author] Cui , Hengjian H. , Li , Runze R. Zhong , Wei W. ( 2015 ). Model-free feature screening for ultrahigh dimensional discriminant analysis . Journal of the American Statistical Association 110 630--641 . barticle

    work page 2015

Show all 59 references
  1. [9]

    Lv , Jinchi J

    barticle [author] Fan , Jianqing J. Lv , Jinchi J. ( 2008 ). Sure independence screening for ultrahigh dimensional feature space . Journal of the Royal Statistical Society: Series B (Statistical Methodology) 70 849--911 . barticle

    2008

  2. [10]

    , Samworth , Richard R

    barticle [author] Fan , Jianqing J. , Samworth , Richard R. Wu , Yichao Y. ( 2009 ). Ultrahigh dimensional feature selection: beyond the linear model . The Journal of Machine Learning Research 10 2013--2038 . barticle

    2009

  3. [11]

    , West , Bettina B

    barticle [author] Foster , Mary M. , West , Bettina B. Francescucci , Anthony A. ( 2011 ). Exploring social media user segmentation and online brand profiles . Journal of Brand Management 19 4--17 . barticle

    2011

  4. [12]

    , Mhiri , Mohamed M

    binproceedings [author] Frikha , Mohamed M. , Mhiri , Mohamed M. , Zarai , Mounir M. Gargouri , Faiez F. ( 2016 ). Time-sensitive trust calculation between social network friends for personalized recommendation . In proceedings of the 18th annual international conference on el...

    2016

  5. [13]

    , Lin , Tzu-Heng T.-H

    barticle [author] Gao , Chen C. , Lin , Tzu-Heng T.-H. , Li , Nian N. , Jin , Depeng D. Li , Yong Y. ( 2021 ). Cross-platform item recommendation for online social e-commerce . IEEE Transactions on Knowledge and Data Engineering 35 1351-1364 . 10.1109/TKDE.2021.3098702 barticle

    2021

  6. [14]

    , Ren , Haojie H

    barticle [author] Guo , Xu X. , Ren , Haojie H. , Zou , Changliang C. Li , Runze R. ( 2023 ). Threshold selection in feature screening for error rate control . Journal of the American Statistical Association 118 1773--1785 . barticle

    2023

  7. [15]

    Wright , Julian J

    barticle [author] Hagiu , Andrei A. Wright , Julian J. ( 2020 ). When data creates competitive advantage . Harvard business review 98 94--101 . barticle

    2020

  8. [16]

    , Li , Runze R

    barticle [author] Huang , Danyang D. , Li , Runze R. Wang , Hansheng H. ( 2014 ). Feature screening for ultrahigh dimensional categorical data with applications . Journal of Business & Economic Statistics 32 237--244 . barticle

    2014

  9. [17]

    , Guan , Guoyu G

    barticle [author] Huang , Danyang D. , Guan , Guoyu G. , Zhou , Jing J. Wang , Hansheng H. ( 2018 ). Network-based naive Bayes model for social network . Science China Mathematics 61 627--640 . barticle

    2018

  10. [18]

    , Witten , Daniela D

    bbook [author] James , Gareth G. , Witten , Daniela D. , Hastie , Trevor T. Tibshirani , Robert R. ( 2013 ). An introduction to statistical learning 112 . Springer . bbook

    2013

  11. [19]

    , Li , Jialiang J

    barticle [author] Jiang , Binyan B. , Li , Jialiang J. Yao , Qiwei Q. ( 2023 ). Autoregressive networks . Journal of Machine Learning Research 24 1--69 . barticle

    2023

  12. [20]

    , Yu , Zhizhi Z

    barticle [author] Jin , Di D. , Yu , Zhizhi Z. , Jiao , Pengfei P. , Pan , Shirui S. , He , Dongxiao D. , Wu , Jia J. , Yu , Philip P. Zhang , Weixiong W. ( 2021 ). A survey of community detection approaches: From statistical modeling to deep learning . IEEE Transactions on Kn...

    2021

  13. [21]

    , Zubcsek , Peter Pal P

    barticle [author] Katona , Zsolt Z. , Zubcsek , Peter Pal P. P. Sarvary , Miklos M. ( 2011 ). Network effects and personal influences: The diffusion of an online social network . Journal of Marketing Research 48 425--443 . barticle

    2011

  14. [22]

    barticle [author] Kipf , Thomas N T. N. Welling , Max M. ( 2016 ). Semi-supervised classification with graph convolutional networks . arXiv preprint arXiv:1609.02907 . barticle

    2016 arXiv

  15. [23]

    , Marmer , Vadim V

    barticle [author] Kojevnikov , Denis D. , Marmer , Vadim V. Song , Kyungchul K. ( 2021 ). Limit theorems for network dependent random variables . Journal of Econometrics 222 882--908 . barticle

    2021

  16. [24]

    , Qu , Xi X

    bbook [author] Lewbel , Arthur A. , Qu , Xi X. , Tang , Xun X. et al. ( 2021 ). Social Networks with Mismeasured Links . Boston College . bbook

    2021

  17. [25]

    , Levina , Elizaveta E

    barticle [author] Li , Tianxi T. , Levina , Elizaveta E. Zhu , Ji J. ( 2019 ). Prediction models for network-linked data . The Annals of Applied Statistics 13 132--164 . barticle

    2019

  18. [26]

    Xu , Chen C

    barticle [author] Li , Xingxiang X. Xu , Chen C. ( 2024 ). Feature screening with conditional rank utility for big-data classification . Journal of the American Statistical Association 119 1385--1395 . barticle

    2024

  19. [27]

    , Zhong , Wei W

    barticle [author] Li , Runze R. , Zhong , Wei W. Zhu , Liping L. ( 2012 ). Feature screening via distance correlation learning . Journal of the American Statistical Association 107 1129--1139 . barticle

    2012

  20. [28]

    , Peng , Heng H

    barticle [author] Li , Gaorong G. , Peng , Heng H. , Zhang , Jun J. Zhu , Lixing L. ( 2012 ). Robust rank correlation based screening . The Annals of Statistics 40 1846--1877 . barticle

    2012

  21. [29]

    , Qian , Cheng C

    barticle [author] Li , Tianxi T. , Qian , Cheng C. , Levina , Elizaveta E. Zhu , Ji J. ( 2020 ). High-dimensional Gaussian graphical models on network-linked data. Journal of Machine Learning Research 21 74--1 . barticle

    2020

  22. [30]

    , Chen , Haojie H

    barticle [author] Liu , Jie J. , Chen , Haojie H. Yang , Yang Y. ( 2022 ). Prediction models with graph kernel regularization for network data . Journal of Applied Statistics 1--18 . barticle

    2022

  23. [31]

    , Zhong , Wei W

    barticle [author] Liu , JingYuan J. , Zhong , Wei W. Li , RunZe R. ( 2015 ). A selective overview of feature screening for ultrahigh-dimensional data . Science China Mathematics 58 1--22 . barticle

    2015

  24. [32]

    , Ke , Yuan Y

    barticle [author] Liu , Wanjun W. , Ke , Yuan Y. , Liu , Jingyuan J. Li , Runze R. ( 2020 ). Model-free feature screening and FDR control with Knockoff features . Journal of the American Statistical Association 0 1--16 . barticle

    2020

  25. [33]

    Vayatis , Nicolas N

    barticle [author] Lugosi , G \'a bor G. Vayatis , Nicolas N. ( 2004 ). On the Bayes-risk consistency of regularized boosting methods . The Annals of statistics 32 30--55 . barticle

    2004

  26. [34]

    , Tang , Weijing W

    barticle [author] Ma , Jiaqi J. , Tang , Weijing W. , Zhu , Ji J. Mei , Qiaozhu Q. ( 2019 ). A flexible generative framework for graph-based semi-supervised learning . Advances in Neural Information Processing Systems 32 3281--3290 . barticle

    2019

  27. [35]

    Zou , Hui H

    barticle [author] Mai , Qing Q. Zou , Hui H. ( 2013 ). The Kolmogorov filter for variable screening in high-dimensional binary classification . Biometrika 100 229--234 . barticle

    2013

  28. [36]

    Zou , Hui H

    barticle [author] Mai , Qing Q. Zou , Hui H. ( 2015 ). The fused Kolmogorov filter: A nonparametric model-free screening method . The Annals of Statistics 43 1471--1497 . barticle

    2015

  29. [37]

    , Abuhashesh , M M

    barticle [author] Maqableh , M M. , Abuhashesh , M M. , Dahabiyeh , L L. , Nawayseh , M M. Masadeh , R R. ( 2021 ). The effect of Facebook users’ satisfaction and trust on stickiness: the role of perceived values . International Journal of Data and Network Science 5 245--256 ....

    2021

  30. [38]

    , Carroll , Micah M

    barticle [author] Milli , Smitha S. , Carroll , Micah M. , Wang , Yike Y. , Pandey , Sashrika S. , Zhao , Sebastian S. Dragan , Anca D A. D. ( 2025 ). Engagement, user satisfaction, and the amplification of divisive content on social media . PNAS nexus 4 pgaf062 . barticle

    2025

  31. [39]

    , Niu , Ziang Z

    barticle [author] Mukherjee , Somabha S. , Niu , Ziang Z. , Halder , Sagnik S. , Bhattacharya , Bhaswar B B. B. Michailidis , George G. ( 2021 ). High dimensional logistic regression under network dependence . arXiv preprint arXiv:2110.03200 . barticle

    2021 arXiv

  32. [40]

    Wang , HanSheng H

    barticle [author] Pan , Rui R. Wang , HanSheng H. ( 2015 ). A note on testing conditional independence for social network analysis . Science China Mathematics 58 1179--1190 . barticle

    2015

  33. [41]

    , Wang , Xueqin X

    barticle [author] Pan , Wenliang W. , Wang , Xueqin X. , Zhang , Heping H. , Zhu , Hongtu H. Zhu , Jin J. ( 2019 a). Ball covariance: A generic measure of dependence in banach space . Journal of the American Statistical Association 529 307--317 . barticle

    2019

  34. [42]

    , Wang , Xueqin X

    barticle [author] Pan , Wenliang W. , Wang , Xueqin X. , Xiao , Weinan W. Zhu , Hongtu H. ( 2019 b). A generic sure independence screening procedure . Journal of the American Statistical Association 114 928--937 . barticle

    2019

  35. [43]

    , Ting , Kai Ming K

    barticle [author] Pang , Ming M. , Ting , Kai Ming K. M. , Zhao , Peng P. Zhou , Zhi-Hua Z.-H. ( 2020 ). Improving deep forest by screening . IEEE Transactions on Knowledge and Data Engineering 34 4298--4312 . barticle

    2020

  36. [44]

    , Ying , Z Z

    barticle [author] Sit , T T. , Ying , Z Z. Yu , Yi Y. ( 2021 ). Event history analysis of dynamic networks . Biometrika 108 223--230 . barticle

    2021

  37. [45]

    Wu , Ying Y

    binproceedings [author] Su , Bing B. Wu , Ying Y. ( 2019 ). Learning distance for sequences by learning a ground metric . In International Conference on Machine Learning 6015--6025 . PMLR . binproceedings

    2019

  38. [46]

    , Cai , Zhanrui Z

    barticle [author] Tong , Zhaoxue Z. , Cai , Zhanrui Z. , Yang , Songshan S. Li , Runze R. ( 2023 ). Model-free conditional feature screening with FDR control . Journal of the American Statistical Association 118 2575--2587 . barticle

    2023

  39. [47]

    , Aribarg , Anocha A

    barticle [author] Wang , Jing J. , Aribarg , Anocha A. Atchad \'e , Yves F Y. F. ( 2013 ). Modeling choice interdependence in a social network . Marketing Science 32 977--997 . barticle

    2013

  40. [48]

    , Faust , Katherine K

    bbook [author] Wasserman , Stanley S. , Faust , Katherine K. et al. ( 1994 ). Social network analysis: Methods and applications . Cambridge university press . bbook

    1994

  41. [49]

    Leng , Chenlei C

    barticle [author] Wu , Weichi W. Leng , Chenlei C. ( 2023 ). A Random Graph-based Autoregressive Model for Networked Time Series . arXiv preprint arXiv:2309.08488 . barticle

    2023 arXiv

  42. [50]

    , Lin , Yuanyuan Y

    barticle [author] Xie , Jinhan J. , Lin , Yuanyuan Y. , Yan , Xiaodong X. Tang , Niansheng N. ( 2020 ). Category-adaptive variable screening for ultra-high dimensional heterogeneous categorical data . Journal of the American Statistical Association 115 747--760 . barticle

    2020

  43. [51]

    , Jiang , Binyan B

    barticle [author] Yan , Ting T. , Jiang , Binyan B. , Fienberg , Stephen E S. E. Leng , Chenlei C. ( 2019 ). Statistical inference in a directed network model with covariates . Journal of the American Statistical Association 114 857--868 . barticle

    2019

  44. [52]

    , Li , Ting T

    barticle [author] Yu , Xianshi X. , Li , Ting T. , Ying , Ningchen N. Jing , Bing-Yi B.-Y. ( 2021 ). Collaborative Filtering with Awareness of Social Networks . Journal of Business & Economic Statistics 1--13 . barticle

    2021

  45. [53]

    , Xu , Gongjun G

    barticle [author] Zhang , Xuefei X. , Xu , Gongjun G. Zhu , Ji J. ( 2022 ). Joint latent space models for network data with high-dimensional node variables . Biometrika 109 707--720 . barticle

    2022

  46. [54]

    Zhu , Liping L

    barticle [author] Zhang , Yaowu Y. Zhu , Liping L. ( 2024 ). Projective independence tests in high dimensions: the curses and the cures . Biometrika 111 1013--1027 . barticle

    2024

  47. [55]

    , Guo , Lingyun L

    barticle [author] Zhang , Mingli M. , Guo , Lingyun L. , Hu , Mu M. Liu , Wenhua W. ( 2017 ). Influence of customer engagement with company social networks on stickiness: Mediating effect of customer value creation . International Journal of Information Management 37 229--240 ...

    2017

  48. [56]

    , Pan , Rui R

    barticle [author] Zhang , Xu X. , Pan , Rui R. , Guan , Guoyu G. , Zhu , Xuening X. Wang , Hansheng H. ( 2020 ). Logistic regression with network structure . Statistica Sinica 30 673--693 . barticle

    2020

  49. [57]

    , Tian , Yahui Y

    barticle [author] Zhang , Xu X. , Tian , Yahui Y. , Guan , Guoyu G. Gel , Yulia R Y. R. ( 2021 ). Depth-based classification for relational data with multiple attributes . Journal of Multivariate Analysis 184 104732 . barticle

    2021

  50. [58]

    , Qian , Chen C

    barticle [author] Zhong , Wei W. , Qian , Chen C. , Liu , Wanjun W. , Zhu , Liping L. Li , Runze R. ( 2023 ). Feature screening for interval-valued response with application to study association between posted salary and required skills . Journal of the American Statistical As...

    2023

  51. [59]

    , Pan , Rui R

    barticle [author] Zhu , Xuening X. , Pan , Rui R. , Li , Guodong G. , Liu , Yuewen Y. , Wang , Hansheng H. et al. ( 2017 ). Network vector autoregression . The Annals of Statistics 45 1096--1123 . barticle

    2017

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.