DocumentCode :
1964133
Title :
k-Means Has Polynomial Smoothed Complexity
Author :
Arthur, David ; Manthey, Bodo ; Roglin, H.
Author_Institution :
Dept. of Comput. Sci., Stanford Univ., Stanford, CA, USA
fYear :
2009
fDate :
25-27 Oct. 2009
Firstpage :
405
Lastpage :
414
Abstract :
The k-means method is one of the most widely used clustering algorithms, drawing its popularity from its speed in practice. Recently, however, it was shown to have exponential worst-case running time. In order to close the gap between practical performance and theoretical analysis, the k-means method has been studied in the model of smoothed analysis. But even the smoothed analyses so far are unsatisfactory as the bounds are still super-polynomial in the number n of data points. In this paper, we settle the smoothed running time of the k-means method. We show that the smoothed number of iterations is bounded by a polynomial in n and 1/¿, where sigma is the standard deviation of the Gaussian perturbations. This means that if an arbitrary input data set is randomly perturbed, then the k-means method will run in expected polynomial time on that input set.
Keywords :
Gaussian processes; computational complexity; pattern clustering; Gaussian perturbations; k-means method; polynomial smoothed complexity; smoothed analysis; smoothed running time; standard deviation; Application software; Biology; Clustering algorithms; Computer science; Data compression; Information retrieval; Mathematics; Performance analysis; Polynomials; Upper bound; clustering; k-means; smoothed analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Foundations of Computer Science, 2009. FOCS '09. 50th Annual IEEE Symposium on
Conference_Location :
Atlanta, GA
ISSN :
0272-5428
Print_ISBN :
978-1-4244-5116-6
Type :
conf
DOI :
10.1109/FOCS.2009.14
Filename :
5438612
Link To Document :
بازگشت