Optimal construction of k-nearest-neighbor graphs for identifying noisy clusters
From MaRDI portal
Publication:1017654
Abstract: We study clustering algorithms based on neighborhood graphs on a random sample of data points. The question we ask is how such a graph should be constructed in order to obtain optimal clustering results. Which type of neighborhood graph should one choose, mutual k-nearest neighbor or symmetric k-nearest neighbor? What is the optimal parameter k? In our setting, clusters are defined as connected components of the t-level set of the underlying probability distribution. Clusters are said to be identified in the neighborhood graph if connected components in the graph correspond to the true underlying clusters. Using techniques from random geometric graph theory, we prove bounds on the probability that clusters are identified successfully, both in a noise-free and in a noisy setting. Those bounds lead to several conclusions. First, k has to be chosen surprisingly high (rather of the order n than of the order log n) to maximize the probability of cluster identification. Secondly, the major difference between the mutual and the symmetric k-nearest neighbor graph occurs when one attempts to detect the most significant cluster only.
Recommendations
- Cluster Identification in Nearest-Neighbor Graphs
- scientific article; zbMATH DE number 1982187
- How the result of graph clustering methods depends on the construction of the graph
- Optimal clustering of multipartite graphs
- Top-\(k\) and clustering with noisy comparisons
- Optimal hierarchical clustering on a graph
Cites work
- scientific article; zbMATH DE number 3862231 (Why is no real title available?)
- scientific article; zbMATH DE number 3280855 (Why is no real title available?)
- A graph-based estimator of the number of clusters
- Cluster Identification in Nearest-Neighbor Graphs
- Combinatorial methods in density estimation
- Connectivity of the mutual \(k\)-nearest-neighbor graph in clustering and outlier detection
- Consistency of Single Linkage for High-Density Clusters
- Paths in graphs
- Percolation
- Probability Inequalities for Sums of Bounded Random Variables
- Random Geometric Graphs
- Threshold Functions, Node Isolation, and Emergent Lacunae in Sensor Networks
Cited in
(18)- Nearest neighbor clustering: a baseline method for consistent clustering with arbitrary objective functions
- Fully adaptive density-based clustering
- Cluster Identification in Nearest-Neighbor Graphs
- Estimation of level set trees using adaptive partitions
- SELP: semi-supervised evidential label propagation algorithm for graph data clustering
- Level set tree methods
- Statistical analysis of a hierarchical clustering algorithm with outliers
- Unconstrained and curvature-constrained shortest-path distances and their approximation
- On the normalized Laplacian spectra of random geometric graphs
- Spectral clustering based on local linear approximations
- On clustering procedures and nonparametric mixture estimation
- Noise-enhanced clustering and competitive learning algorithms
- A survey of neighborhood construction algorithms for clustering and classifying data points
- Batch edge-removal clustering based on random $k$NN graph
- A Review on Modal Clustering
- The coreness and H-index of random geometric graphs
- scientific article; zbMATH DE number 2089226 (Why is no real title available?)
- Improved spectral convergence rates for graph Laplacians on \(\varepsilon \)-graphs and \(k\)-NN graphs
This page was built for publication: Optimal construction of \(k\)-nearest-neighbor graphs for identifying noisy clusters
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1017654)
