Šeperić, Stela (2014) Kriteriji kompleksnosti za kmeans algoritam. Diploma thesis, Faculty of Science > Department of Mathematics.

PDF
Language: Croatian Download (1MB)  Preview 

Archive
(dodatni materijali)
Language: Croatian Download (2MB) 
Abstract
In this thesis, we have studied method for determining $k$ in kmeans algorithm, which is a well known problem. It is shown that the objective function coverges. Furthermore, a couple of lemmas proved that, if data follow a uniform distribution, then the objective function behaves as $C\cdot\frac{1}{k^2}$. Assumption about data is not as strong as assumption in Gmeans algorithm which requires normal distribution of data. We have shown that with relatively limited assumptions, we can neatly describe behaviour of the objective function with respect to $k$  the number of clusters. Finally, a series of tests on simulated data sets confirmed our theoretical results about behaviour of the objective function.
Item Type:  Thesis (Diploma thesis) 

Supervisor:  Goldstein, Pavle 
Date:  2014 
Number of Pages:  29 
Subjects:  NATURAL SCIENCES > Mathematics 
Divisions:  Faculty of Science > Department of Mathematics 
Depositing User:  Iva Prah 
Date Deposited:  10 Jul 2015 12:37 
Last Modified:  10 Jul 2015 12:37 
URI:  http://digre.pmf.unizg.hr/id/eprint/4113 
Actions (login required)
View Item 