A hierarchical clustering algorithm and an improvement of the single linkage criterion to deal with noise

•This paper proposes a hierarchical clustering algorithm robust to noise.•It includes a single linkage improvement that involves local density.•It forbids the merging of representative clusters.•Performance is assessed using data with known ground truth. Hierarchical clustering is widely used in dat...

Full description

Saved in:
Bibliographic Details
Published in:Expert systems with applications Vol. 128; pp. 96 - 108
Main Authors: Ros, Frédéric, Guillaume, Serge
Format: Journal Article
Language:English
Published: New York Elsevier Ltd 15.08.2019
Elsevier BV
Elsevier
Subjects:
Agglomerative
Algorithms
Cluster analysis
Clustering
Criteria
Data mining
Density
Dissimilarity
Environmental Sciences
Noise
Noise sensitivity
Parameter identification
Parameter sensitivity
Sensitivity analysis
Dissimilarity
Density
Agglomerative
DENSITY
DISSIMILARITY
AGGLOMERATIVE
ISSN:0957-4174, 1873-6793
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•This paper proposes a hierarchical clustering algorithm robust to noise.•It includes a single linkage improvement that involves local density.•It forbids the merging of representative clusters.•Performance is assessed using data with known ground truth. Hierarchical clustering is widely used in data mining. The single linkage criterion is powerful, as it allows for handling various shapes and densities, but it is sensitive to noise11A sample code is available at: http://frederic.rosresearch.free.fr/mydata/homepage/.. Two improvements are proposed in this work to deal with noise. First, the single linkage criterion takes into account the local density to make sure the distance involves core points of each group. Second, the hierarchical algorithm forbids the merging of representative clusters, higher than a minimum size, once identified. The experiments include a sensitivity analysis to the parameters and a comparison of the available criteria using datasets known in the literature. The latter proved that local criteria yield better results than global ones. Then, the three single linkage criteria were compared in more challenging situations that highlighted the complementariness between the two levels of improvement: the criterion and the clustering algorithm.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0957-4174
1873-6793
DOI:10.1016/j.eswa.2019.03.031