Fast K-means clustering using deletion by center displacement and norms product (CDNP)

Authors:
Suiang-Shyan Lee;Ja-Chen Lin
Affiliations:
Department of Computer Science, National Chiao Tung University, Hsinchu, Taiwan 30050;Department of Computer Science, National Chiao Tung University, Hsinchu, Taiwan 30050
Venue:
Pattern Recognition and Image Analysis
Year:
2013

Citing 6
Cited 0

Integrating AHP and data mining for product recommendation based on customer lifetime value

Information and Management
A fast VQ codebook generation algorithm using codeword displacement

Pattern Recognition
A fast k-means clustering algorithm using cluster center displacement

Pattern Recognition
Automatic hierarchical clustering algorithm for remote sensing data

Pattern Recognition and Image Analysis
Least squares quantization in PCM

IEEE Transactions on Information Theory
Fast VQ encoding by an efficient kick-out condition

IEEE Transactions on Circuits and Systems for Video Technology

Quantified Score

Hi-index	0.00

Visualization

Abstract

K-means clustering method has been widely used in many areas. However, it is time-consuming when data are in high dimensional space, or when there are many clusters. We try to accelerate its speed by combing our previous work with the simplest version of a fast K-means method that gracefully used the centers' displacement between two iterations. Experimental results show our method not only is several times faster than the fast K-Means method using center displacement; but also accelerates the fast K-means method that used norms-product test only. As a result, the proposed hybrid method is much faster than the plain K-means method. Hence, it is very useful in real-time data mining; examples include medical diagnostics, customer analysis, and vector quantization.