Quantum clustering in non-spherical data distributions: Finding a suitable number of clusters
| Autor: | Raul V. Casana-Eslava, José D. Martín-Guerrero, Paulo J. G. Lisboa, Ian H. Jarman |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
0301 basic medicine
Clustering high-dimensional data Mathematical optimization Cognitive Neuroscience Single-linkage clustering Correlation clustering 02 engineering and technology Computer Science Applications Hierarchical clustering Determining the number of clusters in a data set 03 medical and health sciences 030104 developmental biology Artificial Intelligence 0202 electrical engineering electronic engineering information engineering Cluster (physics) 020201 artificial intelligence & image processing QA Cluster analysis Algorithm k-medians clustering Mathematics |
| Zdroj: | Neurocomputing. 268:127-141 |
| ISSN: | 0925-2312 |
| DOI: | 10.1016/j.neucom.2017.01.102 |
| Popis: | Quantum Clustering (QC) provides an alternative approach to clustering algorithms, several of which are based on geometric relationships between data points. Instead, QC makes use of quantum mechanics concepts to find structures (clusters) in data sets by finding the minima of a quantum potential. The starting point of QC is a Parzen estimator with a fixed length scale, which significantly affects the final cluster allocation. This dependence on an adjustable parameter is common to other methods. We propose a framework to find suitable values of the length parameter σ by optimising twin measures of cluster separation and consistency for a given cluster number. This is an extension of the Separation and Concordance framework previously introduced for K-means clustering. Experimental results on two synthetic data sets and three challenging real-world data sets show that optimisation of cluster separation identifies QC solutions with consistently high Jaccard score measured against true-cluster labels while optimisation of cluster consistency provides insights into hierarchical cluster structure. © 2017 Elsevier B.V. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|
Full Text from ScienceDirect