Stability of density-based clustering

zbMATH Open1283.62130arXiv1011.2771MaRDI QIDQ5405149FDOQ5405149

Authors: Alessandro Rinaldo, Aarti Singh, Rebecca Nugent, Larry Wasserman 

Publication date: 1 April 2014

Abstract: High density clusters can be characterized by the connected components of a level set $L(lambda)=x:p(x)>lambda$ of the underlying probability density function $p$ generating the data, at some appropriate level $lambdageq0$. The complete hierarchical clustering can be characterized by a cluster tree $$. In this paper, we study the behavior of a density level set estimate $widehatL(lambda)$ and cluster tree estimate $widehatcalT$ based on a kernel density estimator with kernel bandwidth $h$. We define two notions of instability to measure the variability of $widehatL(lambda)$ and $widehatcalT$ as a function of $h$, and investigate the theoretical properties of these instability measures.

Full work available at URL: https://arxiv.org/abs/1011.2771

Recommendations

• Generalized density clustering
• Fully adaptive density-based clustering
• A statistical model of cluster stability
• A Sober Look at Clustering Stability
• Cluster analysis: a further approach based on density estimation.

zbMATH Keywords

clusteringmodel selectiondensity estimationlevel setsstability

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05)

Cited In (23)

• Concentration inequalities for empirical processes of linear time series
• Skeleton Clustering: Dimension-Free Density-Aided Clustering
• Mixture model modal clustering
• Measuring the stability of spectral clustering
• Distance Metrics and Clustering Methods for Mixed‐type Data
• Classification methods for Hilbert data based on surrogate density
• Nonparametric inference via bootstrapping the debiased estimator
• A conformal prediction approach to explore functional data
• Nonparametric estimation of regression level sets using kernel plug-in estimator
• The Modal Age of Statistics
• Solution manifold and its statistical applications
• Moving Up the Cluster Tree with the Gradient Flow
• DBSCAN: optimal rates for density-based cluster estimation
• Data stability in clustering: a closer look
• Single linkage clustering and continuum percolation
• Generalized cluster trees and singular measures
• Confidence regions for level sets
• Empirical evolution equations
• Visual stability analysis for model selection in graded possibilistic clustering
• Fully adaptive density-based clustering
• Gromov-Hausdorff stability of linkage-based hierarchical clustering methods
• Generalized density clustering
• A population background for nonparametric density-based clustering

This page was built for publication: Stability of density-based clustering

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5405149)