Spectral Clustering Based on Analysis of Eigenvector Properties

Abstract : In this paper we propose a new method for choosing the number of clusters and the most appropriate eigenvectors, that allow to obtain the optimal clustering. To accomplish the task we suggest to examine carefully properties of adjacency matrix eigenvectors: their weak localization as well as the sign of their values. The algorithm has only one parameter — the number of mutual neighbors. We compare our method to several clustering solutions using different types of datasets. The experiments demonstrate that our method outperforms in most cases many other clustering algorithms.
Document type :
Conference papers
Khalid Saeed; Václav Snášel. 13th IFIP International Conference on Computer Information Systems and Industrial Management (CISIM), Nov 2014, Ho Chi Minh City, Vietnam. Springer, Lecture Notes in Computer Science, LNCS-8838, pp.43-54, 2014, Computer Information Systems and Industrial Management. 〈10.1007/978-3-662-45237-0_6〉
Liste complète des métadonnées

Cited literature [17 references]  Display  Hide  Download

https://hal.inria.fr/hal-01405553
Contributor : Hal Ifip <>
Submitted on : Wednesday, November 30, 2016 - 10:46:20 AM
Last modification on : Monday, January 22, 2018 - 11:08:07 AM
Document(s) archivé(s) le : Monday, March 27, 2017 - 7:48:21 AM

File

978-3-662-45237-0_6_Chapter.pd...
Files produced by the author(s)

Licence


Distributed under a Creative Commons Attribution 4.0 International License

Identifiers

Citation

Małgorzata Lucińska, Sławomir Wierzchoń. Spectral Clustering Based on Analysis of Eigenvector Properties. Khalid Saeed; Václav Snášel. 13th IFIP International Conference on Computer Information Systems and Industrial Management (CISIM), Nov 2014, Ho Chi Minh City, Vietnam. Springer, Lecture Notes in Computer Science, LNCS-8838, pp.43-54, 2014, Computer Information Systems and Industrial Management. 〈10.1007/978-3-662-45237-0_6〉. 〈hal-01405553〉

Share

Metrics

Record views

41

Files downloads

424