Clustering is a data mining unsupervised learning technique used to place data elements
into related groups without advance knowledge of the group
definitions. One of the most popular and widely studied
clustering methods that minimize the clustering error for
points in Euclidean space is called K-means clustering.
However, the k-means method converges to one of many local
minima, and it is known that the final results depend on the
initial starting points (means). In this research paper, we have
introduced and tested an improved algorithm to start the k-
means with good starting points (means). The good initial
starting points allow k-means to converge to a better local
minimum; also the numbers of iteration over the full dataset
are being decreased. Experimental results show that initial
starting points lead to good solution reducing the number of
iterations to form a cluster.
%0 Journal Article
%1 lkulkarni2011iterative
%A Madhuri A Dalal, Nareshkumar D Harale
%A Kulkarni, Umesh L
%B 3
%D 2011
%E Das, Dr. Vinu V
%I The Association of Copmuter Electronics and Electrical Engineers
%J International Journal on Network Security
%K clustering clustering_algorithms data_mining k-means_clustering
%N 3
%P 4
%T An Iterative Improved k-means Clustering
%U http://doi.searchdl.org/01.IJNS.2.3.183
%V 2
%X Clustering is a data mining unsupervised learning technique used to place data elements
into related groups without advance knowledge of the group
definitions. One of the most popular and widely studied
clustering methods that minimize the clustering error for
points in Euclidean space is called K-means clustering.
However, the k-means method converges to one of many local
minima, and it is known that the final results depend on the
initial starting points (means). In this research paper, we have
introduced and tested an improved algorithm to start the k-
means with good starting points (means). The good initial
starting points allow k-means to converge to a better local
minimum; also the numbers of iteration over the full dataset
are being decreased. Experimental results show that initial
starting points lead to good solution reducing the number of
iterations to form a cluster.
@article{lkulkarni2011iterative,
abstract = {Clustering is a data mining unsupervised learning technique used to place data elements
into related groups without advance knowledge of the group
definitions. One of the most popular and widely studied
clustering methods that minimize the clustering error for
points in Euclidean space is called K-means clustering.
However, the k-means method converges to one of many local
minima, and it is known that the final results depend on the
initial starting points (means). In this research paper, we have
introduced and tested an improved algorithm to start the k-
means with good starting points (means). The good initial
starting points allow k-means to converge to a better local
minimum; also the numbers of iteration over the full dataset
are being decreased. Experimental results show that initial
starting points lead to good solution reducing the number of
iterations to form a cluster.
},
added-at = {2012-09-25T09:06:05.000+0200},
author = {Madhuri A Dalal, Nareshkumar D Harale and Kulkarni, Umesh L},
biburl = {https://www.bibsonomy.org/bibtex/2fde5d1aca9a6c43bac66854200d40dc8/ideseditor},
editor = {Das, Dr. Vinu V},
interhash = {c9cec5d1720571f9df1db7b65f0d5c1e},
intrahash = {fde5d1aca9a6c43bac66854200d40dc8},
journal = {International Journal on Network Security},
keywords = {clustering clustering_algorithms data_mining k-means_clustering},
month = {July},
number = 3,
organization = {Institute of Doctors Engineers and scientist(IDES)},
pages = 4,
publisher = {The Association of Copmuter Electronics and Electrical Engineers},
series = 3,
timestamp = {2012-09-25T09:06:06.000+0200},
title = {An Iterative Improved k-means Clustering
},
url = {http://doi.searchdl.org/01.IJNS.2.3.183},
volume = 2,
year = 2011
}