A Highly Efficient Fast Global K-Means Clustering Algorithm

DOI

10.2991/cmes-15.2015.166How to use a DOI?

Keywords

clustering algorithm; k-means; global k-means; fast global k-means

Abstract

To improve clustering effects of fast global K-means and reduce time complexity, a highly efficient fast global K-means algorithm is proposed in this paper. Which, maximal point of density in data sets is chosen as the first initial clustering center, then finding the next initial clustering center, firstly we can exclude a certain number of clusters around the given clustering center ,and narrow selection range of the next initial clustering center, further utilize the related theorem of triangle inequality ,reduce computational amount ,choose the sample which have great contribution in reducing error sum of squares and are apart from the given clustering center as the next initial clustering center, then the modified fast global K-means algorithm reassigns sample to cluster, which will collect sample to the weighted distance of cluster center ,and partition the sample to cluster when weighted distance is minimum. The modified algorithm can select more reasonable initial cluster center, obtain more objective ,true clustering results and shorten the clustering time . The experiment results shows the algorithm is valid.

Copyright

© 2015, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Xian Liang
AU  - Fuheng Qu
AU  - Yong Yang
AU  - Hua Cai
PY  - 2015/04
DA  - 2015/04
TI  - A Highly Efficient Fast Global K-Means Clustering Algorithm
BT  - Proceedings of the 2nd International Conference on Civil, Materials and Environmental Sciences
PB  - Atlantis Press
SP  - 609
EP  - 612
SN  - 2352-5401
UR  - https://doi.org/10.2991/cmes-15.2015.166
DO  - 10.2991/cmes-15.2015.166
ID  - Liang2015/04
ER  -