Pith. sign in

REVIEW 3 major objections 5 minor 35 references

Exponentially Consistent Nonparametric Linkage-Based Clustering of Data Sequences

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read The paper proves that single-linkage clustering of i.i.d. data sequences from unknown distributions is exponentially consistent whenever $d_I < d_H$, where $d_I$ is the largest, over all partitions of a true cluster into two parts, of the…

desk verdict The main MMD proof for SLINK under d_I < d_H is sound and the sequential extension is real; fix the missing KSD proof and the d_I <= d_L typo. read the letter →

arxiv 2411.13922 v4 pith:BAK4XBZW submitted 2024-11-21 stat.ML cs.ITcs.LGeess.SPmath.IT

Bhupender Singh , Ananth Ram Rajagopalan , Srikrishna Bhashyam This is my paper · ORCID
classification stat.MLcs.ITcs.LGeess.SPmath.IT MSC 62H3062L10
keywords nonparametricclusteringdatasequencessingle-linkagesequentialexponentialconsistencymaximummeandiscrepancyKolmogorov-SmirnovdistanceSLINK-SEQ
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 establishes that single-linkage (SLINK) clustering of i.i.d. data sequences generated from unknown distributions is exponentially consistent under the separation condition $d_I < d_H$, where $d_I$ is the maximum distance between the two parts of any partition of a true cluster and $d_H$ is the minimum distance between distinct true clusters. This is weaker than the previously required $d_L < d_H$, with $d_L$ the maximum intra-cluster distance, so the result covers clusters with large diameter that are separated by small gaps, a setting where $k$-medoids clustering can fail. The proof uses estimated maximum mean discrepancy (MMD) distances and a concentration inequality for the estimator to show the error probability decays as $a_f e^{-b_f n}$. The paper further proposes a sequential version, SLINK-SEQ, and proves that its error probability decays exponentially with the expected number of samples: $E[N] \le -\frac{1}{\alpha}\log P_{\max}(1+o(1))$.

What carries the argument

The load-bearing object is the cluster-splitting distance $d_I$: for each true cluster, take every partition of its distributions into two nonempty parts, record the smallest distance between the parts, and let $d_I$ be the largest such value over all clusters. The single-linkage merging rule always joins the two current clusters separated by the smallest estimated distance, and the condition $d_I < d_H$ ensures that at every merge step the closest pair of current clusters lies inside one true cluster. The proof's workhorse is the concentration inequality from [20, Thm. 7], which bounds $P(|\hat{d}(i,j,n)-d(p_i,p_j)| > 4\sqrt{G/n}+\epsilon) \le 2\exp(-n\epsilon^2/(4G))$ for the MMD estimator; Lemma 1 and Lemma 2 convert this into exponential bounds on the events $\hat{d}_H < d_{th}$ and $\hat{d}_I > d_{th}$, giving the rate $b_f = (d_H-d_I)^2/(64G)$. SLINK-SEQ adds a stopping rule that compares the minimum inter-cluster distance of the current clustering with the threshold $C/\sqrt{n}$.

What would settle it

Run fixed-sample-size SLINK on a configuration with known $d_I < d_H$ (for instance the Gaussian Example 1 of the paper) and estimate $P_e$ for increasing $n$; if the magnitude of the slope of $\ln P_e$ versus $n$ is smaller than $(d_H-d_I)^2/(64G)$ for infinitely many $n$, the upper bound of Theorem 1 is violated and the claimed exponential rate is wrong.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is Theorem 1: if $d_I < d_H$, then the probability that fixed-sample-size SLINK returns the wrong partition of the $M$ sequences is upper bounded by $P_e \le a_f e^{-b_f n}$ for large $n$. The quantity $d_I$ measures the largest gap that must be crossed to split any true cluster into two nonempty parts; the condition says that every such intra-cluster split is closer than the nearest inter-cluster pair. Consequently SLINK's nearest-neighbour merging finishes every true cluster before it can merge across clusters. The authors extend the same argument to a sequential algorithm SLINK-SEQ, proving almost-sure finite stopping, universal consistency as the threshold constant grows, and exponential consistency in the sense $E[N] \le -\frac{1}{\alpha}\log P_{\max}(1+o(1))$.

Load-bearing premise

The load-bearing premise is a concentration inequality saying that each estimated pairwise distance is exponentially unlikely to deviate from its true value by a fixed margin; if that bound fails or has a slower rate for the estimator actually used, the exponential consistency proofs collapse.

Editorial extensions

If this is right

  • Fixed-sample-size SLINK is guaranteed to return the true clusters with probability of error at most $a_f e^{-b_f n}$ whenever $d_I < d_H$, so clusters with large internal diameter and small inter-cluster gaps are no longer excluded.
  • The result extends the class of distribution-clustering problems known to be exponentially solvable by a linkage algorithm, since $d_I < d_H$ can hold while the earlier condition $d_L < d_H$ fails; for CLINK the same proof collapses to $d_I = d_L$, recovering the old condition.
  • The sequential algorithm SLINK-SEQ stops almost surely, and its error probability satisfies $E[N] \le -\frac{1}{\alpha}\log P_{\max}(1+o(1))$, so the expected stopping time scales with $1/d_H^2$ while the error probability decays exponentially in the stopping time.
  • In simulations on Gaussian, Gaussian-mixture, MNIST, and MovieLens data, SLINK-SEQ reaches a given probability of error with fewer expected samples than fixed-sample-size SLINK, and the improvement is observed for both MMD- and KSD-based distances.

Reading between the lines

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

  • If the concentration inequality holds for the biased MMD estimator exactly as used, the same $d_I < d_H$ argument should transfer to any distance between distributions whose empirical estimate concentrates exponentially, broadening the method beyond MMD and KSD.
  • The $d_I$ quantity is a property of a cluster's internal geometry rather than its diameter, which suggests linkage-based clustering is especially suited to elongated or size-imbalanced clusters where centroid-based algorithms fail.
  • A direct testable extension is to replace the threshold $C/\sqrt{n}$ with an adaptive or estimated threshold; the proof indicates the error exponent would still be governed by the gap $(d_H-d_I)^2$, and simulations could check whether the empirical optimum remains $\alpha=1/2$.
  • For overlapping clusters, the partition-based definition of $d_I$ no longer applies; extending the condition to soft partitions would determine whether the exponential consistency result survives outside the disjoint-cluster setting.
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

3 major / 5 minor

Summary. The paper studies fixed-sample-size (FSS) and sequential clustering of i.i.d. data sequences generated from unknown distributions. Its main theoretical contribution is a new separation condition, d_I < d_H, where d_I is the maximum, over bipartitions of a true cluster, of the minimal distance between the two parts, and it proves that FSS single-linkage clustering (SLINK) using MMD estimates is exponentially consistent under this condition, which is weaker than the previously used d_L < d_H. It then proposes a sequential variant, SLINK-SEQ, with a C/√n stopping threshold, and claims universal consistency and exponential consistency in the expected stopping time. The paper includes simulations on Gaussian, Gaussian-mixture, MNIST, and MovieLens data comparing SLINK to k-medoids and SLINK-SEQ to FSS SLINK.

Significance. The d_I < d_H condition is a genuine conceptual weakening of d_L < d_H because d_I can be much smaller than the cluster diameter, so the paper enlarges the class of distribution-clustering problems for which SLINK is provably exponentially consistent. The deterministic argument in Section III-B is correct, and the probabilistic proof for the biased MMD estimator is sound provided the cited concentration result in [20, Thm. 7] holds as stated; the authors also provide code and reproducible simulation settings. The main weaknesses are that the KSD version is asserted rather than proved, Theorem 3's universality statement overclaims the assumptions used in its proof, and one step in the sequential proof needs a union bound to be valid. These are fixable without affecting the central MMD result, which I find convincing.

major comments (3)
  1. [Theorem 3, Section IV-A] The statement that SLINK-SEQ is universally consistent 'under any configuration of the true clusters' is not supported by the proof. The proof begins from Assumption 1 (d_I < d_H) and requires δ with d_I < (1−δ)^2 d_H, so the theorem as proved is restricted to configurations satisfying d_I < d_H. The unrestricted claim is also false: if d_I > d_H, even SLINK with exact distances can merge two true clusters before completing an intra-cluster merge (for example, a cluster with a large internal gap and a nearby cluster inside that gap), so the error probability need not go to zero as C→∞. Please restate the theorem with the d_I < d_H assumption, or clarify that 'universal' means only that the distributions are unknown and nonparametric.
  2. [Theorem 3 proof, around Eq. (11)] The inequality P[Γ_n > T_n, E_n] ≤ P[\hat d(i,j,n) > T_n] is not valid as written, because the pair (i,j) is selected after seeing the event and can depend on the outcome. The correct argument is to take a union bound over all same-cluster pairs whose true distance is at most d_I; the existence of at least one such pair with estimated distance above T_n follows from the definition of d_I when E_n occurs. This introduces only an O(M^2) factor, which does not change the exponential decay in C^2. Please add this union-bound step explicitly.
  3. [Sections III-C, VI and Theorems 1, 3–5] The formal results are proved only for the MMD estimator, while the abstract and conclusions claim exponential consistency for SLINK 'with MMD or KSD distance measures.' The sentence before Theorem 1 says a similar proof can be given for KSD using concentration results in [19], but no KSD concentration lemma or proof is supplied, and Lemmas 1–2 concern MMD only. As written, the KSD claims are assertions rather than theorems. Either restrict the formal claims to MMD or add the KSD concentration bound and adapt Lemmas 1–2 and the sequential theorems accordingly.
minor comments (5)
  1. [Algorithm 1, line 5] The loop condition 'while Γn < C√n' contradicts the stopping threshold T_n = C/√n used in Section IV and in all proofs; with the printed condition the threshold grows with n and the algorithm need not stop. The condition should be 'Γn < C/√n'.
  2. [Section II-D and contribution bullet] The claim that d_I < d_L whenever there are more than two distinct points in a cluster is false for an equilateral three-point configuration, where d_I = d_L. The statement should be qualified, e.g., 'in general' or 'for configurations without ties in the bipartition minima.'
  3. [Appendix A, Lemma 2] The hypothesis states MMD(p_j, p_{j'}) < d_I, but the same concentration bound holds when this distance equals d_I, and the proof of Theorem 1 may need the non-strict version when the minimizing pair in a partition attains the maximum exactly. Please relax the inequality to ≤ d_I.
  4. [Theorem 1 proof and Eq. (8)] In the bound for P[\hat d_I > d_{th}], the sentence 'Here i,j are chosen such that d(i,j,n) < d_I' should refer to the true distance and should be '≤ d_I'; the display in Eq. (8) also appears to be missing parentheses, since the derivation leading to n > ((C+8√G)/((1−δ)d_H))^2 treats (d_H − T_n)/2 as a single term.
  5. [Section V] The text contains a typo: 'true clusrers' should be 'true clusters.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: Theorem 1 and the SLINK-SEQ consistency proofs derive exponential bounds from the independent MMD concentration inequality of Gretton et al. and from the definitional separation condition dI < dH; threshold parameters are free algorithm constants, not fitted to make the theorem true.

full rationale

The paper's central claim, Theorem 1, is derived in a self-contained way: the error event is contained in the union of the events that some inter-cluster MMD estimate falls below a freely chosen threshold dth with dI < dth < dH, or that some dI-defining intra-cluster pair estimate exceeds dth. Both probabilities are bounded by the concentration inequality cited from the external reference [20, Thm. 7] for the biased MMD V-statistic, giving exponents bH = (dH - dth)^2/(16G) and bI = (dth - dI)^2/(16G). No parameter is fitted to data, and dth is an arbitrary analytical slack variable (Remark 1 even fixes dth = (dI + dH)/2). The deterministic part of the argument (if dIhat < dHhat then SLINK merges only within true clusters) follows directly from the definitions of dI and dH and the single-linkage merge rule. The sequential results use only the FSS bound, the same concentration inequality, and standard uniform-integrability arguments adapted from [24, 26]; the self-citation to [26] concerns technique, not the target result, and is not load-bearing circularity. The KSD analogue is asserted rather than proved, and the remark that dI < dL 'in general' is factually false for an equilateral three-point cluster (where dI = dL), but these are correctness or completeness concerns, not circularity. The theorem does not reduce to its inputs by definition, and no fitted quantity is renamed as a prediction.

Assumptions & free parameters 2 free parameters · 6 assumptions · 0 invented entities

The central claim rests on four domain assumptions (i.i.d. sequences, bounded kernel, d_I < d_H separation, known K) and two external analytic inputs: the MMD concentration inequality from [20] and an asserted analogous bound for KSD. No new physical or algorithmic entities are invented. The free parameters are the threshold C and the simulation exponent α; neither is fitted to data in the proofs.

free parameters (2)
  • C (SLINK-SEQ threshold constant) = chosen in simulations; proof requires C > C_M
    Defines the stopping threshold T_n = C/√n. It is a user-selected parameter of the algorithm, not fitted to data for the theorem: exponential consistency holds for all sufficiently large C. The proof of Theorem 3 requires C > C_M, where C_M depends on unknown d_I and d_H.
  • α (threshold exponent in simulations) = 1/2
    In Section V-C, the threshold C/n^α is compared for α=1/3, 1/2, 1, and α=1/2 is observed to perform best. The theory only analyzes α=1/2, so this is a simulation hyperparameter, not a fitted constant in the proofs.
assumptions (6)
  • domain assumption The true distribution clusters satisfy d_I < d_H.
    Assumption 1 in Section IV-A. This is the separation condition on the unknown distributions that all theorems require. It is weaker than d_L < d_H but is still an assumption about the problem instance.
  • standard math The MMD concentration inequality of [20, Thm. 7] holds for the biased V-statistic estimator (2).
    Used in Lemmas 1 and 2 and Theorems 1-3 as the only probabilistic tail bound. The paper cites [20] and does not reproduce the proof; the application to the V-statistic with diagonal terms is assumed.
  • domain assumption Each data sequence is i.i.d. from its unknown distribution.
    Stated in Section II-A. Required for the concentration of MMD and KSD estimates; the entire framework breaks without it.
  • domain assumption The kernel satisfies 0 ≤ k(x,y) ≤ G.
    Section II-C. Standard boundedness condition for the MMD concentration inequality and for the bound on the estimator used in the proofs.
  • domain assumption The number of clusters K is known.
    SLINK and SLINK-SEQ take K as input (Algorithm 1). The paper notes that a lower bound on d_H would suffice if K is unknown, but all proofs assume known K.
  • ad hoc to paper The KSD version of the results follows by an analogous proof using concentration results in [19].
    Stated at the end of Section III-C without derivation. The theorem statements cover both MMD and KSD, but only the MMD proof is provided in the manuscript.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exponentially Consistent Nonparametric Linkage-Based Clustering of Data Sequences." pith.science (2026). https://pith.science/paper/BAK4XBZW

@misc{pith2026241113922,
  author       = {Pith},
  title        = {Pith review of: Exponentially Consistent Nonparametric Linkage-Based Clustering of Data Sequences},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BAK4XBZW}},
  note         = {Machine review of arXiv:2411.13922}
}
abstract

In this paper, we consider nonparametric clustering of $M$ independent and identically distributed (i.i.d.) data sequences generated from {\em unknown} distributions. The distributions of the $M$ data sequences belong to $K$ underlying distribution clusters. Existing results on exponentially consistent nonparametric clustering algorithms, like single linkage-based (SLINK) clustering and $k$-medoids distribution clustering, assume that the maximum intra-cluster distance ($d_L$) is smaller than the minimum inter-cluster distance ($d_H$). First, in the fixed sample size (FSS) setting, we show that exponential consistency can be achieved for SLINK clustering under a less strict assumption, $d_I < d_H$, where $d_I$ is the maximum distance between any two sub-clusters of a cluster that partition the cluster. Note that $d_I < d_L$ in general. Thus, our results show that SLINK is exponentially consistent for a larger class of problems than previously known. In our simulations, we also identify examples where $k$-medoids clustering is unable to find the true clusters, but SLINK is exponentially consistent. Then, we propose a sequential clustering algorithm, named SLINK-SEQ, based on SLINK and prove that it is also exponentially consistent. Simulation results show that the SLINK-SEQ algorithm requires fewer expected number of samples than the FSS SLINK algorithm for the same probability of error.

Figures

Figures reproduced from arXiv: 2411.13922 by the authors.

Figure 1
Figure 1. Illustration of dI for a cluster with 3 distributions. The pairwise distances are d1, d2 and d3 with d1 < d2 < d3. The three ways to partition this cluster into 2 are shown. The corresponding minimum distance between these sub-clusters is also indicated: d2 in one case and d1 in the other cases. Overall, the dI for this cluster is d2. The overall dI for the problem will be the maximum of the dI ’s over all the clust… view at source ↗
Figure 2
Figure 2. Example 1: One big cluster and one small cluster, the means of the [PITH_FULL_IMAGE:figures/full_fig_p009_2.png] view at source ↗
Figure 3
Figure 3. Example 2: Two equal sized clusters, the means of the data sequences [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (9 more)
Figure 4
Figure 4. Figure 4: Example 3: 5 clusters, the means of the data sequences are shown, [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: FSS performance comparison of SLINK (using MMD or KSD as [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: FSS performance comparison of SLINK and KMed for Example 2. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: FSS performance comparison of SLINK, KMed and CLINK for [PITH_FULL_IMAGE:figures/full_fig_p011_7.png]
Figure 9
Figure 9. Figure 9: Performance improvement using SLINK-SEQ compared to SLINK [PITH_FULL_IMAGE:figures/full_fig_p011_9.png]
Figure 10
Figure 10. Figure 10: Performance improvement using SLINK-SEQ compared to SLINK [PITH_FULL_IMAGE:figures/full_fig_p012_10.png]
Figure 13
Figure 13. Figure 13: Comparison of SLINK-SEQ performance for different choice of [PITH_FULL_IMAGE:figures/full_fig_p012_13.png]
Figure 11
Figure 11. Figure 11: Performance improvement using SLINK-SEQ compared to SLINK [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]
Figure 14
Figure 14. Figure 14: Comparison of SLINK-SEQ performance for different choice of [PITH_FULL_IMAGE:figures/full_fig_p013_14.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

35 extracted references · 28 canonical work pages

  1. [28]

    On sequence clustering and supervised dimensionality reduc- tion,

    T. Wang, “On sequence clustering and supervised dimensionality reduc- tion,” Ph.D. dissertation, Syracuse University, 2020

    work page 2020

  2. [19]

    K-medoids clustering of data sequences with composite distributions,

    T. Wang, Q. Li, D. J. Bucci, Y . Liang, B. Chen, and P. K. Varshney, “K-medoids clustering of data sequences with composite distributions,” IEEE Transactions on Signal Processing, vol. 67, no. 8, pp. 2093–2106, 2019

    work page 2019

  3. [26]

    Nonparametric sequential clustering of data streams with com- posite distributions,

    ——, “Nonparametric sequential clustering of data streams with com- posite distributions,” Signal Processing, vol. 204, p. 108827, 2023

    work page 2023

  4. [1]

    Exponentially consistent non- parametric clustering of data streams with composite distributions,

    B. Singh, A. Ram, and S. Bhashyam, “Exponentially consistent non- parametric clustering of data streams with composite distributions,” in ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) , 2024, pp. 9401–9405

    work page 2024

  5. [2]

    Sakurai, L

    Y . Sakurai, L. Li, R. Chong, and C. Faloutsos, Efficient Distribution Mining and Classification , pp. 632–643. [Online]. Available: https: //epubs.siam.org/doi/abs/10.1137/1.9781611972788.58

    work page doi:10.1137/1.9781611972788.58

  6. [3]

    Optimal clustering with bandit feedback,

    J. Yang, Z. Zhong, and V . Y . F. Tan, “Optimal clustering with bandit feedback,” Journal of Machine Learning Research, vol. 25, no. 186, pp. 1–54, 2024. [Online]. Available: http://jmlr.org/papers/v25/22-1088.html

    work page 2024

  7. [4]

    Fast k-means algorithm based on a level histogram for image retrieval,

    C.-H. Lin, C.-C. Chen, H.-L. Lee, and J.-R. Liao, “Fast k-means algorithm based on a level histogram for image retrieval,” Expert Systems with Applications , vol. 41, no. 7, pp. 3276–3283, 2014. [Online]. Available: https://www.sciencedirect.com/science/article/pii/ S0957417413009299

    work page 2014

  8. [5]

    Outlying sequence detection in large data sets: A data-driven approach,

    A. Tajer, V . V . Veeravalli, and H. V . Poor, “Outlying sequence detection in large data sets: A data-driven approach,” IEEE Signal Processing Magazine, vol. 31, no. 5, pp. 44–56, 2014

    work page 2014

Show all 35 references
  1. [6]

    Clustering with bregman divergences,

    A. Banerjee, S. Merugu, I. S. Dhillon, and J. Ghosh, “Clustering with bregman divergences,” Journal of Machine Learning Research , vol. 6, no. 58, pp. 1705–1749, 2005. [Online]. Available: http: //jmlr.org/papers/v6/banerjee05b.html

    2005

  2. [7]

    Clustering by means of medoids,

    L. Kaufmann and P. Rousseeuw, “Clustering by means of medoids,” Data Analysis based on the L1-Norm and Related Methods , pp. 405– 416, 01 1987

    1987

  3. [8]

    Minimum spanning trees and single linkage cluster analysis,

    J. C. Gower and G. J. Ross, “Minimum spanning trees and single linkage cluster analysis,” Journal of the Royal Statistical Society: Series C (Applied Statistics) , vol. 18, no. 1, pp. 54–64, 1969

    1969

  4. [9]

    Hierarchical clustering schemes,

    S. C. Johnson, “Hierarchical clustering schemes,” Psychometrika, vol. 32, no. 3, pp. 241–254, 1967

    1967

  5. [10]

    Shalev-Shwartz and S

    S. Shalev-Shwartz and S. Ben-David, Understanding Machine Learning: From Theory to Algorithms . Cambridge University Press, 2014

    2014

  6. [11]

    Hierarchical clustering given confidence intervals of metric distances,

    W. Huang and A. Ribeiro, “Hierarchical clustering given confidence intervals of metric distances,” IEEE Transactions on Signal Processing , vol. 66, no. 10, pp. 2600–2615, 2018

    2018

  7. [12]

    C. M. Bishop, Pattern recognition and machine learning . Springer, vol. 4, no. 4

  8. [13]

    Data clustering: application and trends,

    G. J. Oyewole and G. A. Thopil, “Data clustering: application and trends,” Artificial Intelligence Review , vol. 56, no. 7, pp. 6439–6475, 2023

    2023

  9. [14]

    Data stream clustering: A survey,

    J. A. Silva, E. R. Faria, R. C. Barros, E. R. Hruschka, A. C. P. L. F. d. Carvalho, and J. a. Gama, “Data stream clustering: A survey,” ACM Comput. Surv. , vol. 46, no. 1, Jul. 2013. [Online]. Available: https://doi.org/10.1145/2522968.2522981

    2013

  10. [15]

    Data stream clustering: a review,

    A. Zubaro ˘glu and V . Atalay, “Data stream clustering: a review,” Artif. Intell. Rev. , vol. 54, no. 2, p. 1201–1236, Feb. 2021. [Online]. Available: https://doi.org/10.1007/s10462-020-09874-x

    2021 doi

  11. [16]

    A systematic review of density grid-based clustering for data streams,

    M. Tareq, E. A. Sundararajan, A. Harwood, and A. A. Bakar, “A systematic review of density grid-based clustering for data streams,” IEEE Access, vol. 10, pp. 579–596, 2022

    2022

  12. [17]

    A novel streaming data clustering algorithm based on fitness proportionate sharing,

    X. Yan, M. Razeghi-Jahromi, A. Homaifar, B. A. Erol, A. Girma, and E. Tunstel, “A novel streaming data clustering algorithm based on fitness proportionate sharing,” IEEE Access, vol. 7, pp. 184 985–185 000, 2019

    2019

  13. [18]

    Finding and tracking multi-density clusters in online dynamic data streams,

    C. Fahy and S. Yang, “Finding and tracking multi-density clusters in online dynamic data streams,” IEEE Transactions on Big Data , vol. 8, no. 1, pp. 178–192, 2022

    2022

  14. [20]

    A kernel two-sample test,

    A. Gretton, K. M. Borgwardt, M. J. Rasch, B. Sch ¨olkopf, and A. Smola, “A kernel two-sample test,” Journal of Machine Learning Research, vol. 13, no. 25, pp. 723–773, 2012. [Online]. Available: http://jmlr.org/papers/v13/gretton12a.html

    2012

  15. [21]

    On the empirical estimation of integral probability metrics,

    B. K. Sriperumbudur, K. Fukumizu, A. Gretton, B. Sch ¨olkopf, and G. R. G. Lanckriet, “On the empirical estimation of integral probability metrics,” Electronic Journal of Statistics , vol. 6, no. none, pp. 1550 – 1599, 2012. [Online]. Available: https://doi.org/10.1214/12-EJS722

    2012 doi

  16. [22]

    On exponentially consistency of linkage- based hierarchical clustering algorithm using kolmogrov-smirnov dis- tance,

    T. Wang, Y . Liu, and B. Chen, “On exponentially consistency of linkage- based hierarchical clustering algorithm using kolmogrov-smirnov dis- tance,” in ICASSP 2020-2020 IEEE International Conference on Acous- tics, Speech and Signal Processing (ICASSP) . IEEE, 2020, pp. 3997– 4001

    2020

  17. [23]

    Linear-complexity exponentially- consistent tests for universal outlying sequence detection,

    Y . Bu, S. Zou, and V . V . Veeravalli, “Linear-complexity exponentially- consistent tests for universal outlying sequence detection,” IEEE Trans- actions on Signal Processing , vol. 67, no. 8, pp. 2115–2128, 2019

    2019

  18. [24]

    Universal sequential outlier hypothesis testing,

    Y . Li, S. Nitinawarat, and V . V . Veeravalli, “Universal sequential outlier hypothesis testing,” Sequential Analysis , vol. 36, no. 3, pp. 309–344, 2017. [Online]. Available: https://doi.org/10.1080/07474946. 2017.1360086

    2017

  19. [25]

    Sequential nonparametric detection of anomalous data streams,

    S. C. Sreenivasan and S. Bhashyam, “Sequential nonparametric detection of anomalous data streams,” IEEE Signal Processing Letters , vol. 28, pp. 932–936, 2021

    2021

  20. [27]

    Maxgap bandit: Adaptive algorithms for approximate ranking,

    S. Katariya, A. Tripathy, and R. Nowak, “Maxgap bandit: Adaptive algorithms for approximate ranking,” in Advances in Neural Informa- tion Processing Systems , H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch ´e-Buc, E. Fox, and R. Garnett, Eds., vol. 32. Curran Associates,...

    2019

  21. [29]

    Scalable clustering algorithms for big data: A review,

    M. A. Mahdi, K. M. Hosny, and I. Elhenawy, “Scalable clustering algorithms for big data: A review,” IEEE Access , vol. 9, pp. 80 015– 80 027, 2021

    2021

  22. [30]

    Learning from distributions via support measure machines,

    K. Muandet, K. Fukumizu, F. Dinuzzo, and B. Sch ¨olkopf, “Learning from distributions via support measure machines,” in Proceedings of the 26th International Conference on Neural Information Processing Systems - Volume 1 , ser. NIPS’12. Red Hook, NY , USA: Curran Associates In...

    2012

  23. [31]

    Gradient-based learning applied to document recognition,

    Y . Lecun, L. Bottou, Y . Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE , vol. 86, no. 11, pp. 2278–2324, 1998

    1998

  24. [32]

    Second workshop on infor- mation heterogeneity and fusion in recommender systems (hetrec2011),

    I. Cantador, P. Brusilovsky, and T. Kuflik, “Second workshop on infor- mation heterogeneity and fusion in recommender systems (hetrec2011),” in Proceedings of the fifth ACM conference on Recommender systems , 2011, pp. 387–388

    2011

  25. [33]

    Almost cost-free communication in federated best arm identification,

    S. R. Kota, P. N. Karthik, and V . Y . F. Tan, “Almost cost-free communication in federated best arm identification,” Proceedings of the AAAI Conference on Artificial Intelligence , vol. 37, no. 7, pp. 8378–8385, Jun. 2023. [Online]. Available: https://ojs.aaai.org/index. php/...

    2023

  26. [34]

    Sequential controlled sensing for composite multihypothesis testing,

    A. Deshmukh, V . V . Veeravalli, and S. Bhashyam, “Sequential controlled sensing for composite multihypothesis testing,” Sequential Analysis, vol. 40, no. 2, pp. 259–289, 2021. [Online]. Available: https://doi.org/10.1080/07474946.2021.1912525

    2021

  27. [35]

    Sequential multi-hypothesis testing in multi-armed bandit problems: An approach for asymptotic optimality,

    G. R. Prabhu, S. Bhashyam, A. Gopalan, and R. Sundaresan, “Sequential multi-hypothesis testing in multi-armed bandit problems: An approach for asymptotic optimality,” IEEE Transactions on Information Theory , vol. 68, no. 7, pp. 4790–4817, 2022

    2022

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.