DocumentCode
1962349
Title
The behavior of k-Means: An empirical study
Author
Javed, Kashif ; Babri, Haroon A. ; Saeed, Mehreen
Author_Institution
Dept. of Electr. Eng., Univ. of Eng. & Technol., Lahore
fYear
2008
fDate
25-26 March 2008
Firstpage
1
Lastpage
6
Abstract
In this paper, we study the behavior of the typical k-Means clustering algorithm by investigating the distributions of the final centroids, the sum-of-squares error and the iterations to convergence. This behavior is observed on two different synthetic data sets. It is found that when the clusters are well isolated from each other, the spread of the solutions found by k-Means algorithm indicates a much larger number of local minima as compared to the data set in which clusters overlap.
Keywords
iterative methods; pattern clustering; statistical distributions; final centroid; k-means clustering algorithm; local minima; sum-of-squares error; synthetic data sets; Clustering algorithms; Computer errors; Computer science; Data analysis; Distributed computing; Euclidean distance; Image coding; Image segmentation; Iterative algorithms; Partitioning algorithms;
fLanguage
English
Publisher
ieee
Conference_Titel
Electrical Engineering, 2008. ICEE 2008. Second International Conference on
Conference_Location
Lahore
Print_ISBN
978-1-4244-2292-0
Electronic_ISBN
978-1-4244-2293-7
Type
conf
DOI
10.1109/ICEE.2008.4553948
Filename
4553948
Link To Document