DocumentCode :
3124058
Title :
Two-phases clustering algorithm based on subtractive clustering and k-nearest neighbors
Author :
Horng-Lin Shieh ; Cheng-Chien Kuo ; Fu-Hsien Chen
Author_Institution :
Dept. of Electr. Eng., St. John´s Univ., Taipei, Taiwan
Volume :
04
fYear :
2013
fDate :
14-17 July 2013
Firstpage :
1802
Lastpage :
1806
Abstract :
In this paper, a hybrid clustering method integrated subtractive clustering (SC) and shared nearest neighbor algorithms is proposed for data clustering. In the SC algorithm the parameter used to determine the radius of each cluster affects the performances of clustering results. This paper induces k-nearest neighbors(k-NN) into SC algorithms to solve mention-above problem. First, this paper evaluates the neighbors of each data. Then, a modified SC algorithm based on k-nearest neighbors is developed for indentifying the cluster centers. Three experiments show that proposed method outperforms fuzzy c-means algorithm.
Keywords :
pattern clustering; SC algorithms; cluster center indentification; data clustering; fuzzy c-means algorithm; hybrid clustering method; k-NN; k-nearest neighbors; shared nearest neighbor algorithms; subtractive clustering; two-phase clustering algorithm; Abstracts; Artificial intelligence; Clustering algorithms; Cybernetics; Gaussian noise; Pattern recognition; Clustering; K-nearest neighbors (k-NN); Noise; Subtractive clustering;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2013 International Conference on
Conference_Location :
Tianjin
Type :
conf
DOI :
10.1109/ICMLC.2013.6890889
Filename :
6890889
Link To Document :
بازگشت