Robust and Efficient Fuzzy C-Means Clustering Constrained on Flexible Sparsity

Feiping Nie; Jinglin Xu; Junwei Han; Mingliang Xu; Xuelong Li

REVIEW 5 major objections 5 minor 58 references

Robust and Efficient Fuzzy C-Means Clustering Constrained on Flexible Sparsity

T0 review · 5 major / 5 minor · reviewed 2026-08-14 · deepseek-v4-flash

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

Pith's one-line read A ranking function that sorts distances turns the L0-constrained fuzzy clustering problem into a closed-form update, and the resulting REFCMFS algorithm runs in time linear in the number of samples.

desk verdict A workable sparse FCM variant whose main algorithmic idea is clean, but the global-optimality claim is unsupported and the reported gains come partly from label-tuned hyperparameters. read the letter →

arxiv 1908.06699 v4 pith:KS3KIWPQ submitted 2019-08-19 cs.LG stat.ML

Jinglin Xu , Junwei Han , Mingliang Xu , Feiping Nie , Xuelong Li This is my paper · ORCID

classification cs.LGstat.ML

keywords fuzzyc-meansclusteringL21-normlossL0-normconstraintrobustsparsemembershipmatrixrankingfunctionconvergenceanalysisefficiency

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 proposes REFCMFS, a fuzzy c-means variant that replaces the usual squared-error loss with an $L_{2,1}$-norm loss and requires each row of the membership matrix to have exactly $\tilde K$ nonzero entries. The central claim is that the $L_0$ constraint, normally treated as combinatorial and hard, can be absorbed into the objective by sorting each sample's distances to the cluster centers and retaining only the $\tilde K$ closest ones, with the remaining memberships set to zero. The paper argues this yields a closed-form membership update, keeps per-iteration complexity linear in the number of samples, and produces higher clustering accuracy and normalized mutual information than K-Means, FCM, spectral clustering, GMM, and recent sparse fuzzy methods on six public face and object datasets. If correct, the work offers a tractable way to get exactly sparse fuzzy partitions without relaxations.

What carries the argument

The ranking function $\psi$ is the mechanism: for each sample, sort the $c$ distances to the current centroids in ascending order, keep the $\tilde K$ smallest, and assign zero membership to the rest. This makes the $L_0$ constraint part of the objective rather than a separate combinatorial constraint, so the optimal membership update follows from the convexity of $x^r$ on positive reals when $r > 1$. The $L_{2,1}$-norm loss is handled by an iterative reweighted least-squares step with auxiliary weights $s_{ik} = 1/(2\|x_i - b_k\|_2)$, and centroids are updated as weighted means. The combination gives per-iteration complexity $O(nc \cdot \max(d,c))$.

What would settle it

Run Algorithm 1 from many random centroid initializations on the ORL dataset and compare the final objective values; if different initializations settle at different objective values, the claim that the algorithm converges to the global optimum is contradicted. A complementary check is to evaluate the method with $\tilde K$ and $r$ chosen without any access to labels: if its accuracy advantage disappears, the reported improvement rests on label-based parameter selection.

Watch

Extended reading notes

Core claim

The central discovery is that the $L_0$ constraint $\|\alpha_i\|_0 = \tilde K$ on each membership vector does not need to be relaxed or approximated. After fixing the centroids $B$, the per-sample subproblem is solved exactly by a ranking function $\psi$ that sorts the distances $h_{ik} = \|x_i - b_k\|_2$ in ascending order; the optimal membership concentrates on the first $\tilde K$ positions with values $\alpha_{i,\psi(k)} = h_{i,\psi(k)}^{1/(1-r)} / \sum_{s=1}^{\tilde K} h_{i,\psi(s)}^{1/(1-r)}$, and all other memberships are zero. The paper claims that Algorithm 1 monotonically decreases the objective of problem (6) in each iteration and converges to the global optimum.

Load-bearing premise

The reported accuracy gains depend on tuning the two parameters $\tilde K$ and $r$ using the known cluster labels of each test dataset, and in a real clustering task those labels are unavailable, so the gains over other methods could shrink when parameters must be chosen without supervision.

Editorial extensions

If this is right

Exactly sparse fuzzy memberships can be produced without relaxation, so each sample is assigned to only $\tilde K$ clusters with weights that still sum to one, interpolating between hard K-Means ($\tilde K = 1$) and full FCM ($\tilde K = c$).
The per-iteration complexity $O(nc \cdot \max(d,c))$ scales linearly in the number of samples, which puts the method ahead of spectral clustering and GMM-based clustering on large datasets.
The $L_{2,1}$-norm data term gives the objective a built-in robustness property against outliers, which the experiments support by larger gains on datasets with illumination and pose variation.
The ranking-based update is exact for the per-sample subproblem, so the algorithm's monotone decrease follows from the same reweighted $L_{2,1}$ argument used in the convergence proof.

Reading between the lines

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

An extension not pursued here is to let $\tilde K$ vary per cluster or be estimated from the data, since the paper fixes one value per dataset and tunes it by grid search.
The same absorption strategy could transfer to other $L_0$-constrained problems with a separable per-row structure, such as sparse subspace clustering or sparse regression, where current methods typically use convex relaxations.
Comparing REFCMFS with another robust loss, such as a capped $L_1$ term, inside the same ranking framework would separate how much of the gain comes from the robust loss versus the exact sparsity constraint.

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.

Desk Editor's Note

read the letter

Read this if you want a clear example of a useful algorithmic idea wrapped in an overclaim. The core of REFCMFS is simple: each sample's membership vector is forced to have exactly K nonzero entries, which are found by ranking the distances to centroids and solving a small convex subproblem on that support. That is a nice, exact way to handle an L0 constraint, and it genuinely distinguishes the paper from RSFKM's regularized sparsity and Sparse FCM's different framework. The alternating algorithm is easy to implement, and the reported running times suggest it is competitive.

The paper does a decent job situating the method relative to prior fuzzy and hard clustering work, and the complexity analysis is plausible. The experiments cover six standard image datasets and the method generally edges out the baselines, including the authors' own RSFKM.

The soft spots are significant, though. Theorem 1 says the algorithm converges to the global optimum. The proof establishes, at most, monotone decrease of a nonconvex objective. That is not a global optimality result, and the paper gives no argument that every limit point is a global minimizer. FCM-type objectives are well known to have local minima, so this claim should simply be dropped or replaced with a convergence-to-a-stationary-point statement. Relatedly, the L2,1-norm robust loss label is loose: the objective is a weighted sum of L2 distances per sample-cluster pair, not the L2,1 norm of a single residual matrix. The authors should either derive the connection or phrase the loss more carefully.

The bigger empirical concern is the parameter tuning. They search K and r using ACC and NMI on the labeled datasets and then report the best results. That is label leakage and makes the reported gains look better than an unsupervised user would get. The paper should use a validation set or otherwise demonstrate that the method is insensitive to the choice of these parameters in a realistic setting.

None of this kills the method. The ranking-based L0 handling is a real, small contribution, and the empirical results, even discounted for tuning, suggest the approach is promising. But the paper's advertised guarantees are stronger than what is delivered.

This paper is for readers in the fuzzy clustering niche. It deserves a serious referee, but the review should return a request for revision: fix or remove the global optimality claim, reword the L2,1 connection, and redo the parameter sensitivity with a clean protocol.

Referee Report

5 major / 5 minor

Summary. The paper proposes REFCMFS, a fuzzy c-means variant that combines a claimed L2,1-norm robust data term with an L0-norm constraint on the membership matrix. The membership matrix is constrained row-wise so that each sample has exactly Ktilde nonzero memberships, and a ranking-based closed-form update is derived for the membership subproblem. The centroids are updated via an iteratively reweighted least-squares scheme. The paper claims per-iteration complexity O(nc·max(d,c)), monotone decrease of the objective, and convergence to the global optimum, and it reports ACC/NMI improvements over several baselines on six image datasets. The central contributions are the flexible sparsity formulation and the claimed exact handling of the L0 constraint without approximation.

Significance. If the claims were fully established, the ranking-based membership update would be an elegant and computationally attractive way to handle an L0-constrained fuzzy clustering problem, and the linear-in-n complexity would make the method relevant for larger datasets. The paper clearly identifies a meaningful gap in existing FCM variants and proposes a concrete algorithmic mechanism. However, the global-optimality theorem is not supported by the proof, the objective actually decreased by the algorithm is not consistently defined, and the experimental protocol tunes hyperparameters on ground-truth labels before reporting the best results. These issues currently prevent the empirical and theoretical claims from being accepted as stated. No machine-checked proofs or code are provided, so the algorithmic claims rest entirely on the paper's derivations and reported experiments.

major comments (5)

[Section IV-B, Theorem 1] Theorem 1 states that Algorithm 1 'monotonically decreases the objective of the problem (6) in each iteration and converges to the global optimum.' The proof establishes only monotone decrease of a lower-bounded objective; it does not show that the limit is a global minimizer. The objective in (7) is nonconvex in (α,B), and the alternating scheme is a block-coordinate method, so global optimality requires a separate argument that the paper does not provide. Since the abstract and the contribution list in Section I both repeat the 'optimality and convergence' guarantee, this overclaim is load-bearing. Please either prove global convergence with a rigorous argument or restate the theorem as convergence to a local minimum or stationary point.
[Sections III-A and IV-B, Eqs. (7), (16), (21)-(23)] The objective being minimized is inconsistent across the paper. Equation (7) and the α-update use squared distances ‖xi−bk‖^2 (with hik=‖xi−bk‖^2), but the B-update is developed as an iteratively reweighted method for the unsquared objective, and inequalities (21)-(23) prove monotone decrease for ∑α^r‖xi−bk‖_2, not for ∑α^r‖xi−bk‖^2. Thus Algorithm 1 is not shown to decrease the stated objective of problem (7). The authors should define one consistent objective for both subproblems; if the intended robust loss is the L2,1-style sum of Euclidean norms, then the membership update should use hik=‖xi−bk‖_2, and if the intended objective is squared distances, then the centroid update should be the standard weighted mean without the auxiliary sik reweighting.
[Section III-A, Eqs. (6) and (7)] Problem (6) imposes a total sparsity constraint ‖α‖_0=K, while problem (7) imposes a per-row constraint ‖αi‖_0=Ktilde with Ktilde=K/n. These two formulations are not equivalent in general: the total number of nonzero entries is n·Ktilde, which equals K only when K/n is an integer and the constraint is applied uniformly to every row. In the experiments, Ktilde is selected directly per dataset, so the model actually solved is (7), not (6). Theorem 1, however, refers to problem (6). The relation between the two formulations and which problem is actually solved should be clarified and the theorem stated accordingly.
[Section V-A.4] The experimental protocol selects the hyperparameters Ktilde and r by searching over ACC and NMI values computed with ground-truth labels and then reports 'the best clustering results with optimal parameters.' This is an oracle-style selection procedure and does not reflect an unsupervised clustering setting; it can inflate the reported gains of REFCMFS relative to baselines whose hyperparameters may not be tuned on the same test labels. The authors should select parameters on a validation split, or report performance over the parameter grid without using ground-truth labels to pick the reported result.
[Section III-A] The paper repeatedly states that ‖xi−bk‖^2 in the objective 'achieves the robustness by using the L2,1-norm on the similarity between xi and bk,' but no derivation shows that problem (7) is an L2,1-norm loss. A weighted sum of squared Euclidean distances is not an L2,1 norm of a residual matrix. If the robust term is intended to be ∑‖xi−bk‖_2 α^r, it should be written explicitly and the subsequent closed-form updates should be derived for that objective.

minor comments (5)

[Section IV-B, Lemma 1, Eq. (20)] The right-hand side of inequality (20) appears to have a typo: the second term should have denominator 2‖ut‖_2, i.e., ‖ut‖_2 − ‖ut‖_2^2/(2‖ut‖_2), to match the standard inequality and the form used in Eq. (22).
[Algorithm 1] The loop condition 'while obj(t−1)−obj(t) = thresh≤ 10−7' is malformed; it should use an absolute value, e.g., |obj(t−1)−obj(t)| > 10−7 as the continuation criterion.
[Table IV] The caption of Table IV lists USPS along with YaleB and COIL100, but the table reports only YaleB and COIL100; USPS is already reported in Table III. Remove 'USPS' from the caption.
[Figure 1] The caption contains a typo: 'Explaination' should be 'Explanation'.
[Section I] The abbreviation for the related method in [42] is used inconsistently as both RSFKM and RSFCM in the introduction; please unify the notation.

Circularity Check

1 steps flagged · score 4.0 of 10

Optimization derivation is self-contained, but the empirical 'best results' are partially circular because \tilde K and r are tuned on the labeled test sets; Theorem 1's global-optimality claim is unsupported but not circular.

fitted input called prediction [Section V-A.4 (Parameter Setup), Tables II-IV, Section V.B]
"We search the optimal ˜κ in the range of (1,c) with different steps corresponding to different datasets. ... and record the best clustering results with optimal parameters."

The hyperparameters \tilde K and r are selected by grid search to maximize ACC and NMI on the same labeled datasets whose ACC/NMI are then reported as the method's performance. The values in Tables II-IV are therefore the selected maxima of the search, not evaluations of a fixed, untuned REFCMFS model, so the empirical claim of 'more promising performance' is partly constructed by the metric used to measure it. This does not make the optimization derivation circular, but it makes the empirical comparison partially circular.

full rationale

The core derivation in Section III.B is not circular: for fixed B, problem (9) separates per sample; for any fixed K-element support, the Lagrangian solution (13) is the exact minimizer of the convex support-constrained problem, and equation (14) shows the minimal value decreases with each h_{i,ψ(k)}, so selecting the K smallest distances is an exact support rule rather than a restatement of the conclusion. The B-update is a standard reweighted least-squares step, and Lemma 1 from [46] is stated in the paper as a checkable external inequality; it is not a self-citation that smuggles in the paper's result. Thus the main optimization 'predictions' are independent of their inputs. The partial circularity is confined to Section V: \tilde K and r are tuned on the labeled datasets using the same ACC/NMI metrics later reported, and the 'best' results are the upper envelope of the grid. Separately, I flag Theorem 1 (Section IV-B) as an unsupported overclaim: the proof establishes at most monotone decrease of the nonconvex objective (6), not convergence to a global optimum; this is a correctness gap to address in review, not evidence of circularity.

Assumptions & free parameters 2 free parameters · 4 assumptions · 0 invented entities

The central method depends on two tuned hyperparameters and several unproved mathematical assertions. The ranking-based L0 solution is correct for the stated subproblem, but the robust-loss interpretation and the global convergence claim are not established. No new physical or model entities are introduced.

free parameters (2)

\tilde K (nonzero memberships per sample) = ORL 10, Yale 9, COIL20 13, USPS 5, YaleB 3, COIL100 75
Chosen by grid search per dataset to maximize ACC and NMI on the labeled data; not derived from an unsupervised criterion.
r (fuzziness exponent) = 1.1
Grid-searched from 1 to 1.5 with step 0.1 and set to 1.1 for the reported results.

assumptions (4)

standard math Lemma 1 from Nie et al. [46]: for any nonzero vectors u^{t+1}, u^t, ||u^{t+1}||_2 - ||u^{t+1}||_2^2/(2||u^t||_2) <= ||u^t||_2 - ||u^t||_2^2/(2||u^t||_2)
Used without proof in Theorem 1 to show monotone decrease of the centroid update.
standard math The function f(x)=x^a is convex on R++ for a>=1 or a<=0
Cited from Boyd and Vandenberghe and used to claim convexity of the membership subproblem.
domain assumption For fixed B, the optimal support of each membership row is the \tilde K smallest values of h_ik = ||x_i - b_k||^2
Equation (13)-(15) assumes this; the paper says it is obvious rather than proving the subset selection is globally optimal.
domain assumption The objective (6)/(16) is equivalent to an L2,1-norm robust loss via the reweighting s_ik = 1/(2||x_i-b_k||_2)
The paper labels the loss as L2,1-norm but never writes an L2,1 objective; the equivalence is asserted, not derived.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Robust and Efficient Fuzzy C-Means Clustering Constrained on Flexible Sparsity." pith.science (2026). https://pith.science/paper/KS3KIWPQ

@misc{pith2026190806699,
  author       = {Pith},
  title        = {Pith review of: Robust and Efficient Fuzzy C-Means Clustering Constrained on Flexible Sparsity},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KS3KIWPQ}},
  note         = {Machine review of arXiv:1908.06699}
}

abstract

Clustering is an effective technique in data mining to group a set of objects in terms of some attributes. Among various clustering approaches, the family of K-Means algorithms gains popularity due to simplicity and efficiency. However, most of existing K-Means based clustering algorithms cannot deal with outliers well and are difficult to efficiently solve the problem embedded the $L_0$-norm constraint. To address the above issues and improve the performance of clustering significantly, we propose a novel clustering algorithm, named REFCMFS, which develops a $L_{2,1}$-norm robust loss as the data-driven item and imposes a $L_0$-norm constraint on the membership matrix to make the model more robust and sparse flexibly. In particular, REFCMFS designs a new way to simplify and solve the $L_0$-norm constraint without any approximate transformation by absorbing $\|\cdot\|_0$ into the objective function through a ranking function. These improvements not only make REFCMFS efficiently obtain more promising performance but also provide a new tractable and skillful optimization method to solve the problem embedded the $L_0$-norm constraint. Theoretical analyses and extensive experiments on several public datasets demonstrate the effectiveness and rationality of our proposed REFCMFS method.

Figures

Figures reproduced from arXiv: 1908.06699 by the authors.

**Figure 1.** Explaination of performing ψ on hi and αi . For instance, supposed that c = 5, K˜ = 3, and hi = [hi,1, hi,2, hi,3, hi,4, hi,5] = [2.4, 3.5, 0.6, 7.8, 1.9], then ψ(hi ) = [hi,ψ(1), hi,ψ(2), hi,ψ(3), hi,ψ(4), hi,ψ(5)] = [0.6, 1.9, 2.4, 3.5, 7.8]. According to the first K˜ elements of ψ(hi ), i.e., [0.6, 1.9, 2.4], optimizing their corresponding membership values and setting the rest membership values as zeros, there i… view at source ↗

**Figure 2.** Parameters K˜ = K n and r sensitivity analyses of REFCMFS on ORL, Yale, COIL20, USPS, YaleB, and COIL100 datasets according to Clustering Accuracy. NMI = 1 when the two sets of clusters are identical, and NMI= 0 when they are independent. 4) Parameter Setup: There are two parameters K˜ and r in our proposed REFCMFS method. The first one K˜ in problem (7) is utilized to adjust the number of nonzero elements in the me… view at source ↗

**Figure 3.** Parameters K˜ = K n and r sensitivity analyses of REFCMFS on ORL, Yale, COIL20, USPS, YaleB, and COIL100 datasets according to Clustering NMI. TABLE II COMPARISON RESULTS ON ORL AND YALE DATASETS IN TERMS OF ACC AND NMI. Methods ORL Yale ACC NMI ACC NMI K-Means 48.60±1.40 71.28±1.37 42.91±6.18 49.66±3.41 K-Means++ 50.70±3.80 73.57±2.17 37.45±4.97 45.59±3.33 K-Medoids 42.10±2.90 63.03±1.84 37.58±3.64 43.51±3.46 GMM 5… view at source ↗

Figures from the paper (2 more)

**Figure 4.** Figure 4: The membership values of each sample for K-Means, FCM, and REFCMFS on ORL, Yale, COIL20, USPS, and YaleB datasets, respectively. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

**Figure 5.** Figure 5: The convergence curves of our proposed REFCMFS method on ORL, Yale, COIL20, USPS, YaleB, and COIL100 datasets, respectively. [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

58 extracted references · 53 canonical work pages

[1]

Clustering and summarizing protein- protein interaction networks: A survey,

S. S. Bhowmick and B. S. Seah, “Clustering and summarizing protein- protein interaction networks: A survey,” IEEE Transactions on Knowl- edge and Data Engineering , vol. 28, no. 3, pp. 638–658, 2016

work page 2016
[2]

A systematic approach to clustering whole trajectories of mobile objects in road networks,

B. Han, L. Liu, and E. Omiecinski, “A systematic approach to clustering whole trajectories of mobile objects in road networks,” IEEE Transac- tions on Knowledge and Data Engineering , vol. 29, no. 5, pp. 936–949, 2017

work page 2017
[3]

Density-based place clustering using geo-social network data,

D. Wu, J. Shi, and N. Mamoulis, “Density-based place clustering using geo-social network data,” IEEE Transactions on Knowledge and Data Engineering, vol. 30, no. 5, pp. 838–851, 2018

work page 2018
[4]

A survey on clustering techniques in medical diagnosis,

N. Nithya, K. Duraiswamy, and P. Gomathy, “A survey on clustering techniques in medical diagnosis,” International Journal of Computer Science Trends and Technology, vol. 1, no. 2, pp. 17–23, 2013

work page 2013
[5]

Comparing the performance of biomedical clustering methods,

C. Wiwie, J. Baumbach, and R. R ´ottger, “Comparing the performance of biomedical clustering methods,” Nature Methods, vol. 12, no. 11, p. 1033, 2015

work page 2015
[6]

Time series clustering by a robust autoregressive metric with application to air pollution,

P. D’Urso, L. De Giovanni, and R. Massari, “Time series clustering by a robust autoregressive metric with application to air pollution,” Chemometrics and Intelligent Laboratory Systems , vol. 141, pp. 107– 124, 2015. 10 ORL Yale COIL20 USPS YaleB REFCMFS FCM K-Means 0 10 20 30 400 0.1 0.2 0.3 0.35 Number of Clusters Membership Values 10 20 30 400.0245 0.0...

work page 2015
[7]

The global k-means clustering algorithm,

A. Likas, N. Vlassis, and J. J. Verbeek, “The global k-means clustering algorithm,” Pattern Recognition, vol. 36, no. 2, pp. 451–461, 2003

work page 2003
[8]

Modiﬁed global k-means algorithm for minimum sum- of-squares clustering problems,

A. M. Bagirov, “Modiﬁed global k-means algorithm for minimum sum- of-squares clustering problems,” Pattern Recognition, vol. 41, no. 10, pp. 3192–3199, 2008

work page 2008

Show all 58 references

[9]

Fast global k-means clustering using cluster membership and inequality,

J. Z. C. Lai and T. J. Huang, “Fast global k-means clustering using cluster membership and inequality,” Pattern Recognition, vol. 43, no. 5, pp. 1954–1963, 2010

1954
[10]

Fast modiﬁed global k-means algorithm for incremental cluster construction,

A. M. Bagirov, J. Ugon, and D. Webb, “Fast modiﬁed global k-means algorithm for incremental cluster construction,” Pattern Recognition , vol. 44, no. 4, pp. 866–876, 2011

2011
[11]

The minmax k-means clustering algorithm,

G. Tzortzis, A. Likas, and G. Tzortzis, “The minmax k-means clustering algorithm,” Pattern Recognition, vol. 47, no. 7, pp. 2505–2516, 2014

2014
[12]

K-means-based consensus clustering: A uniﬁed view,

J. Wu, H. Liu, H. Xiong, J. Cao, and J. Chen, “K-means-based consensus clustering: A uniﬁed view,” IEEE Transactions on Knowledge and Data Engineering, vol. 27, no. 1, pp. 155–169, 2015

2015
[13]

Opti- mized cartesian k-means,

J. Wang, J. Wang, J. Song, X.-S. Xu, H. T. Shen, and S. Li, “Opti- mized cartesian k-means,” IEEE Transactions on Knowledge and Data Engineering, vol. 27, no. 1, pp. 180–192, 2015

2015
[14]

Group k-means,

J. Wang, S. Yan, Y . Yang, M. S. Kankanhalli, S. Li, and J. Wang, “Group k-means,” arXiv preprint arXiv:1501.00825 , 2015

2015 arXiv
[15]

Robust k-means: a theoretical revisit,

A. Georgogiannis, “Robust k-means: a theoretical revisit,” in Advances in Neural Information Processing Systems , 2016, pp. 2891–2899

2016
[16]

I-k-means-+: An iterative clustering algorithm based on an enhanced version of the k-means,

H. Ismkhan, “I-k-means-+: An iterative clustering algorithm based on an enhanced version of the k-means,” Pattern Recognition, vol. 79, pp. 402–413, 2018

2018
[17]

Similarity-based clustering by left-stochastic matrix factorization,

R. Arora, M. R. Gupta, A. Kapila, and M. Fazel, “Similarity-based clustering by left-stochastic matrix factorization,” Journal of Machine Learning Research, vol. 14, no. 1, pp. 1715–1746, 2013

2013
[18]

K-means clustering is matrix factorization,

C. Bauckhage, “K-means clustering is matrix factorization,” arXiv preprint arXiv:1512.07548, 2015

2015 arXiv
[19]

Two-phase clustering process for outliers detection,

M.-F. Jiang, S.-S. Tseng, and C.-M. Su, “Two-phase clustering process for outliers detection,” Pattern Recognition Letters, vol. 22, no. 6-7, pp. 691–700, 2001

2001
[20]

Discovering cluster-based local outliers,

Z. He, X. Xu, and S. Deng, “Discovering cluster-based local outliers,” Pattern Recognition Letters, vol. 24, no. 9-10, pp. 1641–1650, 2003

2003
[21]

Improving k-means by outlier removal,

V . Hautam ´aki, S. Cherednichenko, I. K ´arkk´ainen, T. Kinnunen, and P. Fr´anti, “Improving k-means by outlier removal,” in SCIA, 2005

2005
[22]

Clustering-based outlier detection method,

S.-y. Jiang and Q.-b. An, “Clustering-based outlier detection method,” in FSKD, 2008, pp. 429–433

2008
[23]

A novel k-means algorithm for clustering and outlier detection,

Y . Zhou, H. Yu, and X. Cai, “A novel k-means algorithm for clustering and outlier detection,” in FITME, 2009

2009
[24]

An outlier detection method based on clustering,

R. Pamula, J. K. Deka, and S. Nandi, “An outlier detection method based on clustering,” in EAIT, 2011

2011
[25]

Initialization of k-modes clustering using outlier detection techniques,

F. Jiang, G. Liu, J. Du, and Y . Sui, “Initialization of k-modes clustering using outlier detection techniques,” Information Sciences, vol. 332, pp. 167–183, 2016

2016
[26]

A novel approach to noise 11 clustering for outlier detection,

F. Rehm, F. Klawonn, and R. Kruse, “A novel approach to noise 11 clustering for outlier detection,” Soft Computing , vol. 11, no. 5, pp. 489–494, 2007

2007
[27]

A new local distance-based outlier detection approach for scattered real-world data,

K. Zhang, M. Hutter, and H. Jin, “A new local distance-based outlier detection approach for scattered real-world data,” in PAKDD, 2009

2009
[28]

On integrated clustering and outlier detection,

L. Ott, L. Pang, F. T. Ramos, and S. Chawla, “On integrated clustering and outlier detection,” in NIPS, 2014

2014
[29]

Non-exhaustive, overlap- ping k-means,

J. J. Whang, I. S. Dhillon, and D. F. Gleich, “Non-exhaustive, overlap- ping k-means,” in ICDM, 2015

2015
[30]

k-means clustering with outlier removal,

G. Gan and M. K.-P. Ng, “k-means clustering with outlier removal,” Pattern Recognition Letters, vol. 90, pp. 8–14, 2017

2017
[31]

Robust clustering algorithm for suppression of outliers [data classiﬁcation applications],

B. S. Y . Lam and Y . Hong, “Robust clustering algorithm for suppression of outliers [data classiﬁcation applications],” in International Symposium on Intelligent Multimedia, Video and Speech Processing, 2004, pp. 691– 694

2004
[32]

Clustering spatiotemporal data: An augmented fuzzy c-means,

H. Izakian, W. Pedrycz, and I. Jamal, “Clustering spatiotemporal data: An augmented fuzzy c-means,” IEEE Transactions on Fuzzy Systems , vol. 21, no. 5, pp. 855–868, 2013

2013
[33]

Generalization rules for the suppressed fuzzy c-means clustering algorithm,

L. Szil ´agyi and S. M. Szil ´agyi, “Generalization rules for the suppressed fuzzy c-means clustering algorithm,” Neurocomputing, vol. 139, pp. 298–309, 2014

2014
[34]

A sparse fuzzy c-means algorithm based on sparse clustering framework,

X. Qiu, Y . Qiu, G. Feng, and P. Li, “A sparse fuzzy c-means algorithm based on sparse clustering framework,” Neurocomputing, vol. 157, pp. 290–295, 2015

2015
[35]

Kernel-based fuzzy c-means clustering algorithm based on genetic algorithm,

Y . Ding and X. Fu, “Kernel-based fuzzy c-means clustering algorithm based on genetic algorithm,” Neurocomputing, vol. 188, pp. 233–238, 2016

2016
[36]

Multivariate fuzzy c-means algorithms with weighting,

B. A. Pimentel and R. M. de Souza, “Multivariate fuzzy c-means algorithms with weighting,” Neurocomputing, vol. 174, pp. 946–965, 2016

2016
[37]

Robust-learning fuzzy c-means clustering algorithm with unknown number of clusters,

M.-S. Yang and Y . Nataliani, “Robust-learning fuzzy c-means clustering algorithm with unknown number of clusters,” Pattern Recognition , vol. 71, pp. 45–59, 2017

2017
[38]

Conditional spatial fuzzy c-means clustering algorithm for segmentation of mri images,

S. K. Adhikari, J. K. Sing, D. K. Basu, and M. Nasipuri, “Conditional spatial fuzzy c-means clustering algorithm for segmentation of mri images,” Applied Soft Computing , vol. 34, pp. 758–769, 2015

2015
[39]

Modiﬁed possibilistic fuzzy c-means algorithms for segmentation of magnetic resonance image,

J. Aparajeeta, P. K. Nanda, and N. Das, “Modiﬁed possibilistic fuzzy c-means algorithms for segmentation of magnetic resonance image,” Applied Soft Computing , vol. 41, pp. 104–119, 2016

2016
[40]

Generalized entropy based possibilistic fuzzy c-means for clustering noisy data and its convergence proof,

S. Askari, N. Montazerin, M. F. Zarandi, and E. Hakimi, “Generalized entropy based possibilistic fuzzy c-means for clustering noisy data and its convergence proof,” Neurocomputing, vol. 219, pp. 186–202, 2017

2017
[41]

Signiﬁcantly fast and robust fuzzy c-means clustering algorithm based on morpho- logical reconstruction and membership ﬁltering,

T. Lei, X. Jia, Y . Zhang, L. He, H. Meng, and A. K. Nandi, “Signiﬁcantly fast and robust fuzzy c-means clustering algorithm based on morpho- logical reconstruction and membership ﬁltering,” IEEE Transactions on Fuzzy Systems, 2018

2018
[42]

Robust and sparse fuzzy k-means clustering

J. Xu, J. Han, K. Xiong, and F. Nie, “Robust and sparse fuzzy k-means clustering.” in IJCAI, 2016

2016
[43]

A convergence theorem for teh fuzzy isodata cluster- ing algorithms,

J. C. Bezdek, “A convergence theorem for teh fuzzy isodata cluster- ing algorithms,” IEEE Transactions on Pattern Analysis and Machine Intelligence, no. 1, pp. 1–8, 1980

1980
[44]

C. M. Bishop, Pattern recognition and machine learning . springer, 2006

2006
[45]

F. R. Chung and F. C. Graham, Spectral graph theory . American Mathematical Soc., 1997, no. 92

1997
[46]

Efﬁcient and robust feature selection via joint 𝓁2,1-norms minimization,

F. Nie, H. Huang, X. Cai, and C. H. Ding, “Efﬁcient and robust feature selection via joint 𝓁2,1-norms minimization,” in NIPS, 2010

2010
[47]

Robust dictionary learning wif capped l1-norm,

W. Jiang, F. Nie, and H. Huang, “Robust dictionary learning wif capped l1-norm,” in IJCAI, 2015

2015
[48]

Boyd and L

S. Boyd and L. Vandenberghe, Convex optimization . Cambridge university press, 2004

2004
[49]

Parameterisation of a stochastic model for human face identiﬁcation,

F. S. Samaria and A. C. Harter, “Parameterisation of a stochastic model for human face identiﬁcation,” in ACV, 1994, pp. 138–142

1994
[50]

Eigenfaces vs. ﬁsherfaces: Recognition using class speciﬁc linear projection,

P. N. Belhumeur, J. P. Hespanha, and D. J. Kriegman, “Eigenfaces vs. ﬁsherfaces: Recognition using class speciﬁc linear projection,” Yale University New Haven United States, Tech. Rep., 1997

1997
[51]

Columbia object image library (coil-20),

S. A. Nene, S. K. Nayar, H. Murase et al. , “Columbia object image library (coil-20),” Columbia University, Tech. Rep., 1996

1996
[52]

A database for handwritten text recognition research,

J. J. Hull, “A database for handwritten text recognition research,” IEEE Transactions on Pattern Analysis and Machine Intelligence , vol. 16, no. 5, pp. 550–554, 1994

1994
[53]

From few to many: Illumination cone models for face recognition under variable lighting and pose,

A. S. Georghiades, P. N. Belhumeur, and D. J. Kriegman, “From few to many: Illumination cone models for face recognition under variable lighting and pose,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 23, no. 6, pp. 643–660, 2001

2001
[54]

Columbia object image library (coil 100). department of comp,

S. Nayar, S. A. Nene, and H. Murase, “Columbia object image library (coil 100). department of comp,” Science, Columbia University, Tech. Rep. CUCS-006-96, 1996

1996
[55]

A tutorial on spectral clustering,

U. V on Luxburg, “A tutorial on spectral clustering,” Statistics and Computing, vol. 17, no. 4, pp. 395–416, 2007

2007
[56]

k-means++: The advantages of careful seeding,

D. Arthur and S. Vassilvitskii, “k-means++: The advantages of careful seeding,” in Annual ACM-SIAM Symposium on Discrete Algorithms , 2007

2007
[57]

A simple and fast algorithm for k-medoids clustering,

H.-S. Park and C.-H. Jun, “A simple and fast algorithm for k-medoids clustering,” Expert systems with applications , vol. 36, no. 2, pp. 3336– 3341, 2009

2009
[58]

Large scale spectral clustering with landmark- based representation,

X. Chen and D. Cai, “Large scale spectral clustering with landmark- based representation,” in AAAI, 2011

2011

Pith tools

Reviewed August 14, 2026 · model on record in the stance chip above.

[1] [1]

Clustering and summarizing protein- protein interaction networks: A survey,

S. S. Bhowmick and B. S. Seah, “Clustering and summarizing protein- protein interaction networks: A survey,” IEEE Transactions on Knowl- edge and Data Engineering , vol. 28, no. 3, pp. 638–658, 2016

work page 2016

[2] [2]

A systematic approach to clustering whole trajectories of mobile objects in road networks,

B. Han, L. Liu, and E. Omiecinski, “A systematic approach to clustering whole trajectories of mobile objects in road networks,” IEEE Transac- tions on Knowledge and Data Engineering , vol. 29, no. 5, pp. 936–949, 2017

work page 2017

[3] [3]

Density-based place clustering using geo-social network data,

D. Wu, J. Shi, and N. Mamoulis, “Density-based place clustering using geo-social network data,” IEEE Transactions on Knowledge and Data Engineering, vol. 30, no. 5, pp. 838–851, 2018

work page 2018

[4] [4]

A survey on clustering techniques in medical diagnosis,

N. Nithya, K. Duraiswamy, and P. Gomathy, “A survey on clustering techniques in medical diagnosis,” International Journal of Computer Science Trends and Technology, vol. 1, no. 2, pp. 17–23, 2013

work page 2013

[5] [5]

Comparing the performance of biomedical clustering methods,

C. Wiwie, J. Baumbach, and R. R ´ottger, “Comparing the performance of biomedical clustering methods,” Nature Methods, vol. 12, no. 11, p. 1033, 2015

work page 2015

[6] [6]

Time series clustering by a robust autoregressive metric with application to air pollution,

P. D’Urso, L. De Giovanni, and R. Massari, “Time series clustering by a robust autoregressive metric with application to air pollution,” Chemometrics and Intelligent Laboratory Systems , vol. 141, pp. 107– 124, 2015. 10 ORL Yale COIL20 USPS YaleB REFCMFS FCM K-Means 0 10 20 30 400 0.1 0.2 0.3 0.35 Number of Clusters Membership Values 10 20 30 400.0245 0.0...

work page 2015

[7] [7]

The global k-means clustering algorithm,

A. Likas, N. Vlassis, and J. J. Verbeek, “The global k-means clustering algorithm,” Pattern Recognition, vol. 36, no. 2, pp. 451–461, 2003

work page 2003

[8] [8]

Modiﬁed global k-means algorithm for minimum sum- of-squares clustering problems,

A. M. Bagirov, “Modiﬁed global k-means algorithm for minimum sum- of-squares clustering problems,” Pattern Recognition, vol. 41, no. 10, pp. 3192–3199, 2008

work page 2008