A Validity Index Based on Connectivity

Author

Saha, Sriparna ; Bandyopadhyay, Sanghamitra

Author_Institution

Machine Intell. Unit, Indian Stat. Inst., Kolkata

fYear

2009

fDate

4-6 Feb. 2009

Firstpage

91

Lastpage

94

Abstract

In this paper we have developed a connectivity based cluster validity index. This validity index is able to detect the number of clusters automatically from data sets having well separated clusters of any shape, size or convexity. The proposed cluster validity index, connect-index, uses the concept of relative neighborhood graph for measuring the amount of "connectedness" of a particular cluster. The proposed connect-index is inspired by the popular Dunn\´s index for measuring the cluster validity. Single linkage clustering algorithm is used as the underlying partitioning technique. The superiority of the proposed validity measure in comparison with Dunn\´s index is shown for four artificial and two real-life data sets.

Keywords

data mining; graph theory; pattern classification; pattern clustering; connect-index; connectivity-based cluster validity index; data mining; data set partitioning technique; relative neighborhood graph; single linkage clustering algorithm; unsupervised classification; Clustering algorithms; Clustering methods; Couplings; Machine intelligence; Particle measurements; Partitioning algorithms; Pattern recognition; Shape; Virtual manufacturing; Validity index; clustering; relative neighborhood graph;

fLanguage

English

Publisher

ieee

Conference_Titel

Advances in Pattern Recognition, 2009. ICAPR '09. Seventh International Conference on

Conference_Location

Kolkata

Print_ISBN

978-1-4244-3335-3

Type

conf

DOI

10.1109/ICAPR.2009.53

Filename

4782749