The elbow statistic: Multiscale clustering statistical significance

Francisco J. Perez-Reche

stat.MLMar 3, 2026

The elbow statistic: Multiscale clustering statistical significance

Authors: Francisco J. Perez-Reche

Organizations: School of Natural and Computing Sciences, University of Aberdeen, Fraser Noble Building, AB24 3UE, United Kingdom

Abstract

Selecting the number of clusters remains a fundamental challenge in unsupervised learning. Existing approaches typically focus on identifying a single "optimal" partition, often overlooking statistically meaningful structure present across multiple resolutions. We introduce ElbowSig, a general inferential framework for assessing clustering structure over a range of resolutions. The method formalizes the elbow heuristic by defining a normalized discrete curvature statistic based on the sequence of within-cluster heterogeneity values, and evaluates its significance relative to a null distribution of unstructured data. This yields hypothesis tests across resolutions, enabling simultaneous inference at multiple clustering scales. We derive the asymptotic behavior of the null statistic in both large-sample and high-dimensional regimes, characterizing its limiting form and variability. Because it depends only on the heterogeneity sequence, ElbowSig is compatible with a wide range of clustering algorithms, including hard, fuzzy, and model-based methods. Experiments on synthetic and real datasets show that the procedure controls Type-I error under unstructured data while providing power to detect multiscale organization, revealing structure that is often missed by single-resolution selection criteria.

Figures & tables

Clustering method	Point-wise heterogeneity $h_{k}(x_{i};\bm{\theta})$	Parameters $\bm{\theta}$
Hard clustering (agglomerative and k-means)	Inertia: $\min_{1\leq j\leq k}\\|x_{i}-\mu_{j}\\|^{2}$	$\displaystyle\mu_{j}$
Fuzzy $c$ –means (FCM)	Weighted inertia: $\displaystyle\sum_{j=1}^{k}w_{ij}^{m}\,\\|x_{i}-\mu_{j}\\|^{2}$	$\displaystyle(\mu_{j},m,\{w_{ij}\})$
Gaussian mixture model (GMM)	Negative log-likelihood: $\displaystyle-\ln\!\Big(\sum_{j=1}^{k}\pi_{j}\,\phi(x_{i}\!\mid\!\mu_{j},\Sigma_{j})\Big)$	$\displaystyle(\pi_{j},\mu_{j},\Sigma_{j})$

Table 1: Point-wise heterogeneity measures

h_{k}(x_{i};\bm{\theta})

for the clustering models considered in this paper. For hard clustering (agglomerative and k-means) and FCM,

\mu_{j}

denotes the centroid of cluster

C_{j}

. For FCM,

w_{ij}\in[0,1]

are the membership weights satisfying

\sum_{j=1}^{k}w_{ij}=1

m>1

is the fuzzifier exponent. For GMM, each component has mixing weight

\pi_{j}

(

\sum_{j=1}^{k}\pi_{j}=1

), mean

\mu_{j}

, and covariance

\Sigma_{j}

;

\phi(x_{i}\mid\mu_{j},\Sigma_{j})

denotes the corresponding Gaussian density.

Figure 1: Mean value and variance of the elbow statistic $\delta_{k}^{(r)}$ for unstructured data as a function of the dimension $D$ for agglomerative, k-means, GMM and FCM clustering methods. The expectation and variance are computed from $N_{R}=500$ reference datasets of size $N=30$ uniformly distributed on $[0,1]^{D}$ , assuming $k=3$ clusters. The dashed lines indicate the predicted asymptotic trends for large $D$ . The dotted line in panel (a) indicates the asymptotic value of $\mathbb{E}[\delta_{k}^{(r)}]$ for FCM in the limit of large fuzzifier parameter $m$ .

Figure 2: Clustering of two-dimensional data. (a) Dataset with $N=200$ observations consisting of $M=3$ Gaussian components centred at the locations indicated by crosses, each with spread $\sigma_{c}=1$ . Solid circles show data points coloured according to their generating component. Grey crosses show one realization of a bounding-box uniform reference dataset. Partitions with $k\in\{1,2,\dots,10\}$ were evaluated using five standard methods for estimating the number of clusters: (b) Calinski–Harabasz, (c) Davies–Bouldin, and (d) silhouette scores as functions of $k$ . Vertical dashed lines indicate the value $\hat{k}$ at which each index attains its maximum. (e) Gap statistic as a function of $k$ , with vertical lines marking the optimal number of clusters according to the Gap (I) and Gap (II) criteria described in the text. Error bars indicate the standard error estimate $s_{k}$ ( Tibshirani et al., 2001 ) .

Figure 3: Application of ElbowSig to the two-dimensional dataset shown in Fig. 2 . (a) Total within-cluster heterogeneity (inertia) as a function of $k$ . (b) Observed elbow statistic (circles) together with the $100(1-p_{\text{sig}})$ percentile of the $\delta_{k}$ distribution obtained from unstructured reference datasets (squares) at level $q_{1}=0.05$ . (c) Empirical p-values $p_{k}$ comparing the observed elbow statistic at each $k$ with its reference distribution. The horizontal dashed line indicates the per-scale significance threshold $p_{\text{sig}}$ .

$D$	$M$	$\sigma_{c}$	Method	Number of clusters, $k$
$D$	$M$	$\sigma_{c}$	Method	1	2	3	4	5	6	7	8	9	10
2	3	1	ElbowSig (per-k)	–	28	82	0	3	13	0	1	5	2
			ElbowSig (FDR)	–	25	80	0	1	11	0	1	5	2
			Gap (I)	6	12	81	1	0	0	0	0	0	0
			Gap (II)	0	8	76	1	2	5	0	1	3	4
			CH	–	16	70	0	1	3	0	1	4	5

Table 2: Results for synthetic clustered datasets with

N=200

observations across dimensions

D\in\{2,5,20\}

, with

M

Gaussian components randomly centred in

[-10,10]^{D}

and within-cluster standard deviation

\sigma_{c}

as indicated. Each entry reports how many of the 100 simulated datasets led a method to select the corresponding number of clusters

k

. Results for the ElbowSig and Gap methods are based on 200 PCA-aligned uniform reference datasets. ElbowSig results are shown for both the individual scale (per-

k

) and global (FDR) significance criteria. Gap (I), Gap (II), Calinski–Harabasz (CH), Davies–Bouldin (DB), and the Silhouette index are included for comparison. Dashes at

k=1

for all methods except the gap-based ones indicate that those methods cannot return

k=1

by definition.

Clust. method	Reference	Control	Number of clusters, $k$
Clust. method	Reference	Control	2	3	4	5	6	7	8	9	10
Agglomerative	BBU	per- $k$	49	100	2	4	43	9	1	10	6
	BBU	FDR	47	100	2	3	41	7	1	8	5
	PCA	per- $k$	2	99	0	0	14	2	1	9	4
	PCA	FDR	2	99	0	0	12	1	0	7	3
K-means	BBU	per- $k$	38	99	14	9	25	5	7	1	3

Table 3: Effect of clustering algorithm, reference dataset construction, and multiple-testing correction on the ElbowSig outcomes. Shown are results from 100 datasets generated from

M=3

Gaussian components with

\sigma_{c}=1

D=5

dimensions. Each realization consists of

N=200

observations using randomly located centres of the Gaussian components in

[-10,10]^{5}

. “BBU” indicates bounding-box uniform reference datasets, and “PCA” denotes PCA-aligned uniform reference datasets. The numbers in each cell give how many datasets led a method to select the corresponding number of clusters

k

D	Data	Method	Reference	Number of clusters, $k$
D	Data	Method	Reference	1	2	3	4	5	6	7	8	9	10
2	Uniform	ElbowSig (per-k)	BBU	(78)	2	5	4	4	0	1	1	1	4
		ElbowSig (FDR)	BBU	(97)	1	0	0	0	0	0	0	0	2
		Gap (I)	BBU	99	1	0	0	0	0	0	0	0	0
		Gap (II)	BBU	11	3	14	19	14	6	9	8	3	13
		SigClust	Gaussian	27	-	-	-	-	-	-	-	-	-

Table 4: Results for unstructured data in dimensions

D\in\{2,20\}

generated from either a Uniform distribution on

[0,1]^{D}

or a Gaussian distribution with variance

\sigma_{c}=1

in each dimension. For each configuration, 100 independent datasets were simulated. Each entry reports how many of the 100 datasets led a method to select the indicated number of clusters. For ElbowSig, the value displayed in parentheses in the

k=1

column denotes the number of datasets for which no statistically significant clustering was detected at any tested value

k>1

. Results are shown for both the individual scale (per-

k

) and global (FDR) significance criteria. Gap (I) and Gap (II) correspond to the two gap-statistic selection rules described in the main text. For SigClust, only the number of cases in which the null hypothesis of a single Gaussian component was not rejected is shown.

Data	$N$	$D$	Classes	Clustering	Reference	$k$ (per- $k$ )	$k$ (FDR)
Iris	150	4	3	Agglom.	BBU	2,3,5	2,3,5
				Agglom.	PCA	3	–
				k-means	BBU	2,3,5	2,3,5
				k-means	PCA	3	3
				GMM	BBU	2,3,5,6,8	2,3,5,6,8
				GMM	PCA	2	2

Table 5: ElbowSig results for real datasets using several clustering algorithms and reference constructions.

Cluster Number Estimation

See all 13 papers

Composite Silhouette: A Subsampling-based Aggregation Strategy
Apr 15, 2026Aggelos Semoglou, Aristidis Likas, John PavlopoulosCluster Number Estimation Clustering
Athens University of Economics and Business, Greece. · Archimedes, Athena Research Center, Greece. · University of Ioannina, Greece.
Selective Inference for Clustering Under Unknown Covariance
Oct 12, 2026Yiqun T. Chen, Lucy L. GaoSelective Inference Statistical Hypothesis Testing
Departments of Biostatistics and Computer Science, Johns Hopkins University · Department of Statistics and Actuarial Science, University of Waterloo
Selective Inference for Deep Clustering in Latent Spaces
Sep 23, 2026Eina Mizui, Tomohiro Shiraishi, Shunichi Nishino +1Selective Inference Clustering
Nagoya University · RIKEN

Apr 15, 2026J/K move · Enter open · S save

Composite Silhouette: A Subsampling-based Aggregation Strategy

Aggelos Semoglou, Aristidis Likas, John Pavlopoulos

Athens University of Economics and Business, Greece. · Archimedes, Athena Research Center, Greece. · University of Ioannina, Greece.

Cluster Number Estimation Clustering Unsupervised Clustering

Open paper AskPDF arXiv

Determining the number of clusters is a central challenge in unsupervised learning, where ground-truth labels are unavailable. The Silhouette coefficient is a widely used internal validation metric for this task, yet its standard micro-averaged form tends to favor larger clusters under size imbalance. Macro-averaging mitigates this bias by weighting clusters equally, but may overemphasize noise from under-represented groups. We introduce Composite Silhouette, an internal criterion for cluster-count selection that aggregates evidence across repeated subsampled clusterings rather than relying on a single partition. For each subsample, micro- and macro-averaged Silhouette scores are combined through an adaptive convex weight determined by their normalized discrepancy and smoothed by a bounded nonlinearity; the final score is then obtained by averaging these subsample-level composites. We establish key properties of the criterion and derive finite-sample concentration guarantees for its subsampling estimate. Experiments on synthetic and real-world datasets show that Composite Silhouette effectively reconciles the strengths of micro- and macro-averaging, yielding more accurate recovery of the ground-truth number of clusters.

Abstract

Cluster Number Estimation

Explore similar work

Composite Silhouette: A Subsampling-based Aggregation Strategy

Selective Inference for Clustering Under Unknown Covariance

Selective Inference for Deep Clustering in Latent Spaces