Search by:
Year of publication
Author name
Paper title
Метод тривимірної тріангуляції в задачах кластерного аналізу
Full text (PDF)
UDC: 519.237
Publication Language: Russian
Stuc. intelekt. 2010; 15(1):16-24
Abstract: The method of multidimensional scaling on the basis of the 3-D triangulation is presented. The qualitative preservation possibility of geometrical structure of objects by multidimensional space mapping to three- dimensional space is considered. The results of application of the method for clustering problem of the Mendeleyev periodic table are presented. It is experimentally shown, that the presented method is more effective by criteria of clustering quality in comparison with the methods of k-averages and the Kohonen neural network.
Keywords:
References:
View full text (PDF)