Zastosowanie przetwarzania równoległego w analizie skupień
The application of the parallel computing in cluster analysis

Summary/Abstract: In the social research data sets which include thousands of variables are encountered more and more often. During the research of the group structure of units we encountered a lot of specific problems. One of them is the possibility of making grouping within acceptable time. This problem arises from two basic sources. One of them is the size and speed of expansion of the data set. The second, is the complexity of numerical algorithms, which are used in cluster analysis. The aim of this paper is to demonstrate the applicability of parallel computing in cluster analysis. The basic methods of multithread-ed algorithms for determining the typical statistical measures and multithreaded algorithm of k-means in cluster analysis will be presented. The practical aspects of the applications will be shown on the basis of simulations in Matlab.

• Details
• Contents